Professional Documents
Culture Documents
Title Page
Introduction
Silicon
Digital Cameras: How a ccd Works
How a Smartphone Knows Up from Down
Cesium
How an Atomic Clock Works
Primer: Nuclear Structure
Uranium
The Hardest Step in Making a Nuclear Bomb
Lead
The Lead‐Acid Battery: A Nineteenth Century Invention for the Twenty-First
Century
Aluminum
Anodizing, or The Beauty of Corrosion
Primer: Waves
Tungsten, Thorium, & Copper
How a Microwave Oven Works
Primer: Electrons, Energy Levels, and Light Emission
Chromium, Helium, & Neon
How a Laser Works
Eight Amazing
Engineering Stories
Using the Elements to Create
Extraordinary Technologies
Bill Hammack, Patrick Ryan, & Nick Ziech
RCA engineers pioneered the use of photoconductive elements to record images. They arranged them in an x-y grid -- the
light gray areas are pixels, connected by a grid of wires. Charge accumulates at the center of each pixel, where the wires cross
at its center. To read the charge on each pixel, pulses in the x and y directions open and close the diodes. As the x (horizontal)
pulse moves from left to right, it consecutively opens the diodes. (Diodes allow only current to flow in only one direction; the
positive voltage of the pulse opens the diode.) The y pulses work the same way. The horizontal scan rate is much faster than the
vertical one. That is, the vertical pulse opens a diode attached to a row, and then the horizontal pulse rapidly zips across that
row, opening each pixel in that row. Then the vertical pulse moves down a row, and the horizontal pulse repeats its motion.
Capacitive Coupling
Capacitors in a circuit stores energy. Now, the word storage implies that capacitors introduce a time-varying element in a circuit.
Current flowing in a circuit with only resistors flows at a constant rate, but introduce a capacitor, and the flow can suddenly start or
stop. While many circuits have capacitors built into them, capacitance also shows up unwanted in electronic systems, causing noise
that distorts signals. For example, two unconnected wires side by side will become “capacitively coupled” if they are close enough.
In digital systems, such as an x-y-readout imaging device, elements become so small that these effects can be large: The mess of wires
creates unpredictable crosstalk that just appears as noise. One can see how unwanted capacitance muddies a clear signal by
looking at a square wave pulse.
a) If there is very little or no capacitive coupling, the square wave is undistorted; b) if the resistance times the capacitance in the line equals the period of the square wave, we get significant distortion; c) if the resistance times capacitance far
exceeds the wave’s period, we get severe distortion, only seeing the spikes where the square wave changes voltage.
A CCD is created by “doping” silicon with a small amount of boron, which effectively adds a positive charge carrier to the
silicon. Engineers then create photosensitive sections within the CCD by adding arsenic to the silicon in between the channel
stops. The arsenic adds a negative charge carrier to that region of the silicon. They cover this surface with a thin layer of
insulating silicon dioxide, and then engineers deposit thin strips of metal, typically aluminum, perpendicular to the channel
stops. Note that a pixel, as shown in the insert, is defined by three electrodes and two channel stops..
Just like the silver halides in film cameras or the x-y grid in early digital devices, light strikes the
surface of the CCD when the picture is taken. It passes through the gaps in the electrodes, allowing the
silicon pixels to build up charge. After exposure, the CCD has captured an image‒the only problem is that
it is stored as charges on the surface of a slab of silicon. To take these groups of stored charges and turn
them into an image requires removing the charges from the pixels.
Recall that a camera has two essential functions: to capture and record an image. Right now, we have
the image stuck as charges in the pixels. The great innovation of the CCD was how it moved the image
from the pixels to the camera’s electronics without using external wires and gates that distort the image.
To record the image, the CCD shifts the built-up charges from row to row until it reaches the bottom,
where a read-out register transfers the charge to the camera’s electronics, which in turn construct an
image.
In a modern CCD, this process transfers the charge with an amazing 99.9995% efficiency; that is, with
very little distortion. If we examine the CCD’s edge, looking down the metal electrodes, we can see how
the CCD traps and moves charges. The CCD is a metal oxide semiconductor (MOS). It consists of a
semiconductor (the silicon) covered first by an insulating layer and then by a thin layer of metallic
electrodes. By varying the voltage placed across this device by the electrodes, we can create a trap for
electrons, which are the charges that make up the image.
The great technological advantage of the CCD, when it was first introduced, was the way it moved a captured charge. Rather
than using wires as in an x-y grid, it instead moves the electrons captured by exposure to light row by row through the solid
silicon. (The next two figures describe in detail how this happens.) Highlighted here is a single row (although all rows move) that
is transferred down the CCD slab until a read-out register at the bottom records the charges.
The CCD slab is a MOS (metal oxide semiconductor), which can be used to trap a charge by varying the voltage. To create a
pixel, we use three of these MOS structures side by side.
A) If we apply no voltage to the MOS, mobile negative charge carriers are distributed throughout the
semiconductor.
B) Applying a negative voltage to the metal moves electrons away from the metal-semiconductor
interface.
C) Applying a highly negative voltage drives electrons deep into the bulk of the semiconductor,
leaving positive charge carriers near the surface.
D) Using three of these MOS structures side by side allows us to create a “trap” for electrons.
By lowering the voltage of the center electrode relative to the sides, we form a region with positive
charge. When the light strikes the silicon, electrons that are trapped in this small potential well flow into
the area near the surface. This creates isolated charges on the surface of the CCD’s silicon, charges that are
located at different points across the CCD’s “grid” and make up the image. Now let’s turn to recording that
image, or the details of getting the charge out of the pixels.
Let’s look at a four-pixel section plus one-third of a fifth pixel. As noted above, when light strikes the
center wells, each pixel stores a charge. We then drop the voltage on the electrode to the right of each
well, allowing the stored charge to migrate to the area with the lowest voltage. We then raise the voltage
on the original well, trapping this charge under the next electrode. All charges have now moved to the
right by one-third of a pixel. We continue this process until all of the charges reach the bottom edge where
the camera’s electronics record the information from each row.
Here’s what happens in detail. In the figure above, the four pixels are shown at time 1, immediately
after exposure. From left to right, the pixels have two charges, three charges, one charge, and four charges
trapped at the semiconductor-metal interface. To move the charge one row to the right, the following
happens: At time 2, the potential in each well immediately to the right of the stored charges drops to the
same voltage as in the well to its left. This causes the charges to move to the right. Then at time 3, the
potential of the well that originally trapped the charges rises (Notice that a charge on the left is moving
from a well not seen in time 1). At time 4, the potential in each well immediately to the right of the stored
charges drops to the same voltage as in the well to its left. The charges once more move to the right. Then
at time 5, the potential of the well that previously trapped the charges rises. At time 6, the potential in
each well immediately to the right of the charges that were just moved and stored drops to the same
voltage as in the well to its left. The charges once more move to the right. Then at time 7, the potential of
the well that previously trapped the charges rises. This completes one “clock” cycle: The charges in the
rows have moved down a row. This continues until all of the charge is removed from the CCD.
This figure shows four pixels from four different rows of a CCD and what happens at seven sequential times to these four
pixels.
It may seem like a cumbersome process. Indeed it can be very slow because it is serial; that is, there is
no skipping or jumping by rows. The first row must be transferred first, then the second, and the third, and
so on. You can observe this row-by-row serial reading of an image in the time you need to wait between
taking pictures with a digital camera. This time delay from the serial motion is the price paid to have no
wires as in an x-y device, where individual pixels can be read. The great virtue of the CCD is that it
captures a clean, clear image.
How a CCD Creates Color
A CCD only detects light intensity, not color. This means a single pixel would measure the combined
intensity of all light colors at the same time. We could use this to make a black-and-white image, but to
make a color image we need to separate the entering light into its three components: the primary colors
red, green, and blue. It would then seem obvious to record a color image using three CCDs.
Left Engineers could use three CCDs to create a color image, but this would be very expensive, so instead they use Right a
single CCD covered with a color filter array. This creates an image that is a mosaic of red, green, and blue sections. Full color is
restored using an algorithm. Shown here is the Bayer Color Filter Array.
Using a prism, we could split the incoming light into three rays and pass one through a red filter, one
through a green filter, and the last through a blue filter. A CCD behind each would record the intensity of
each color, and then we could recombine them into a color image. Although this solution is simple,
engineers rarely use such a design in consumer digital cameras because it’s far too expensive and
requires three CCDs! Instead, they use a bit of math to create the color image. A typical consumer camera
uses a single CCD with a color filter array (CFA) placed in the path of the incoming light. Engineers divide
this array into pixel-sized sections: Some of the sections have a red filter, some green, and some blue.
This means that the image that comes out of the CCD is a mosaic of red, green, and blue sections. The
camera then applies an algorithm to estimate the correct colors and fill in the picture. For example, if a
red filter covers a pixel, we would need to estimate the green and blue components of that pixel. To do
this we’d use the adjacent pixels. For instance, if the intensity of green color in a nearby pixel is 5% and
the intensity of green in another nearby pixel is 7%, a good estimate of the green intensity in the pixel of
interest–the one covered by the red filter–would be 6%. This process works because the image’s
significant details are much larger than the pixels. Sounds impossible, but you’ve seen the results
yourself: Every digital picture you’ve ever snapped likely uses this method!
Active Pixel Sensors (APS)
Your smartphone likely doesn’t use a CCD; it’s probably an APS (active pixel sensor) CMOS chip. In
spite of the CCD’s great advantage over the first x-y pixel devices, the CCD still has disadvantages. This
should be no surprise because every engineered object is a balance or trade-off among desired
properties. The quest to make both larger imaging devices (telescopes, for example) and smaller ones
(cell phones) highlights the downside of CCD technology. Making large CCDs calls for more efficient row-
to-row charge transfer. Every time a transfer occurs, a little bit of charge is lost. In a 1024 x 1024 CCD,
for instance, about 1% of the charge is lost from the last pixel read. (The last pixel is the worse-case
scenario: It must be transferred 1,024 times to the readout register, then down that register another 1,024
steps. For these 2,048 steps, the amount of charge transferred would be 0.99999952048 = 0.989 charge
transferred or about 1.2% lost). In an array that is 8192 x 8192, this same efficiency would mean about
8% would be lost from the farthest pixel. So, to make the 8192 x 8192 CCD have the same amount lost as
the 1024 x 1024 array would require somehow increasing the charge transfer to 99.99993%. One can see
that eventually the size of the CCD will outstrip any attempts to make the efficiency higher. In addition, the
CCD is slow to read because it is discharged row by row. In making tiny cameras for cell phones, the CCD
has two main limitations. First, the CCD has to integrate onto a chip with other components. Second, the
CCD requires large voltages, perhaps 10 to 15 volts, that can drain a cell phone battery. Oddly, the way
forward for imaging devices is to, in a way, return to the x-y devices that lost out to the CCD in the 1970s.
The early devices suffered from severe capacitive coupling that distorted the images recorded. These
early devices were passive pixel sensors (PPS). New chip-making methods allow production of x-y
devices with a transistor built into every pixel. In this APS device, the transistor functions as an amplifier
that increases the signal from a pixel, thus overcoming the noise from capacitance. In addition, the
transistor allows digital filtering techniques to be used to reduce noise, something that could not be done
in the early PPS system.
The CCD and the Nobel Prize
In 2009, the Nobel Committee awarded the Nobel Prize in physics to Willard S. Boyle and George E. Smith for “the invention of
an imaging semiconductor circuit -- the CCD sensor.” In a way, their work on the CCD was inspired by magnetic bubble memories,
which was all the rage in 1969 at Bell Laboratories. These memories used tiny magnetized spots, called bubbles, moved about by
currents to store information. This worried Boyle. As executive director for the division of Bell Labs that worked on silicon -- the
dominant medium for computer memory at the time -- he worried that these new bubbles might divert funding and support from silicon
research. Boyle invited his friend and colleague George Smith to help him come up with a competitor to this new technology. On a
chalkboard, they devised a way to use silicon, silicon dioxide, and metal electrodes to store charge in specific areas on the surface
of the silicon. The conversation took about an hour as they jotted down in their notebooks that the device could be used as “an
imaging device” and a “display device” in addition to being used for computer memory. After reflecting on their chalkboard talk for
a few weeks, they decided to build a prototype CCD. Within a week, they had a working device that proved the concept of their idea.
They did not pursue using the CCD to capture images; that task fell to their Bell Labs colleague Michael Tompsett.
A great engineer, Tompsett carefully turned the brainstorm of a charge-coupled device into a reality. His name alone appears on
the first patent for the CCD as an imaging device; the patent was titled, appropriately enough, “Charge transfer image devices. ”
While one might feel Tompsett was overlooked for the Nobel Prize, the prize is generally given for the invention or discovery of
fundamental concepts in physics. Boyle and Smith indeed laid down the idea of a CCD, but controversy arises because the only
practical application of a CCD is as an imaging device.
How a Smartphone Knows Up from Down
HUMANS AND OTHER LIVING THINGS can tell up from down because we feel the pull of gravity.
For instance, if you drop this book, gravity will accelerate it toward the center of the earth until it comes
to rest on the ground. But how does a smartphone know which way the phone is rotated, so the screen is
always right-side up? To ensure that the screen on your smartphone always is pointed in the correct
direction, engineers place tiny accelerometers inside the phone to orient it with respect to the earth. An
accelerometer is a device that measures gravitational pull. But how can these tiny accelerometers tell
which way is up?
Basics of an Accelerometer
You can see the essential principles of an accelerometer in the simple device shown in the next figure.
When that device is upright, gravity stretches the spring downward, as indicated by the mark labeled 1g,
meaning one unit of gravitational acceleration. (One g is what you feel with little or no other acceleration
on earth; for reference, when a roller coaster car takes off you feel about 2.5gs, and fighter pilots black
out at about 10gs.) On the device 0g occurs when the tube lays flat so the spring feels no gravitational
pull, and the spring has no extension. This distance from 0g to 1g sets the scale for marking 2g, 3g, etc.
We can use this accelerometer to measure the upward vertical acceleration experienced by the tube. If
we were to quickly jolt the tube upwards, we would see the weight drop inside the tube, possibly to 2 or
3gs. Three of these basic accelerometers can be used tell us the orientation of an object.
A simple accelerometer: A glass tube with cork stoppers at each end makes up the housing. The seismic mass is a lead ball
tethered to the housing by a spring.
In an oblong box, align one accelerometer each along the x, y and z axes. By measuring changes in the
lengths of the springs, we can detect which edge points up relative to gravity. In the first position, the x-
axis accelerometer records 1g, while in the y and z direction the weights lie against the side of the tubes
and the springs are not extended. Rotate the box so it sits on its long edge, and the y-axis accelerometer
will register 1g, while the x and z accelerometers will read 0g. Although the accelerometer inside a
smartphone is a bit more complex, it works using these same principles.
Three of these simple accelerometers can determine the orientation of a box. Note the change in the x and z
accelerometers. When the box lies on its z-axis, the ball in the accelerometer along that axis lies flat while the spring in the
accelerometer on the x-axis is extended. When the box is rotated, the ball in the x-axis accelerometer lies flat, while the ball
stretches the spring in the accelerometer along the z-axis.
Technology & Privacy
We marvel at the great things technology can do, like detect the subtle motions of a cell phone. As with all technologies, though,
power comes with peril -- even with such a seemingly innocuous object as a smartphone accelerometer. Researchers at the Georgia
Institute of Technology showed that such an accelerometer could track the keystrokes of a computer user. They placed a phone on
the table beside a computer keyboard. Since the accelerometer samples a phone’s vibration about 100 times per second, it would
jiggle slightly with each press of the neighboring keyboard. (The keyboard moved the table slightly, and the ultra-sensitive
accelerometer detected these small vibrations of the table.) By detecting pairs of keystrokes located relative to each other on the
keyboard, they could guess with 80% accuracy the word being typed. For example, in typing “canoe” the user would create four
keystroke pairs: C-A, A-N, N-O, and O-E. The accelerometer would determine for those pairs if the two letters were related in one of
four ways: Left-Left-Near, Left-Right-Far, Right-Right-Far, and Right-Left-Far. They compared this to a preloaded dictionary of
likely words to determine the most probable word made for these pairs of keystrokes. Small wonder, then, that Melvin Kranzberg, a
historian, once observed “technology is neither good nor bad; nor is it neutral.”
The leftmost image shows a two-plate capacitor before the circuit is closed. There are no charges on the two metal plates.
When the circuit is closed (center image) a charge builds up on the plates and current stops flowing. If the plates are moved
closer together, as in the image on the left, current flows until the charge is redistributed on the plates.
A smartphone accelerator doesn’t quite work like this, but we’re getting close. The problem is that our
two-plate capacitor has a serious defect.
In the simple two-plate capacitor, the relationship between plate position and current is non-linear.
For example, if we reduce the distance by 25%, the capacitance will drop by 6%. If we again change the
distance by 25%, the capacitance drops by about 4.5%. This makes the accelerometer difficult to
calibrate. Engineers prefer a linear response; that is, if we reduce the distance by one-quarter and then by
another quarter, we’d like to see the same drop in capacitance each time. This would allow for a more
uniform sensitivity across the useful range. So if we set up a slightly different capacitor where the top and
bottom plates are stationary and the middle one moves, the current generated by changes in this
differential capacitor will be linear.
In this differential capacitor, we measure the difference in the charges that form in the bottom capacitor (the bottom and
middle plate) and the top capacitor (the middle and top plate). On the left, the capacitor is uncharged. If we use the same size
batteries and keep the middle plate exactly between the top and bottom plate, then positive charge (as shown in the middle
image) will build up equally on both plates. If we move the middle plate closer to the bottom plate, the capacitance of the bottom
capacitor increases while the top decreases. The difference in capacitance is linear with respect to the motion of the middle
plate.
Look at what happens if we charge the differential capacitor using two batteries with opposite
polarities: Current flows until the middle plate becomes negative and the top and bottom plates are
positive. Now, if we move the middle plate, current will flow to redistribute the charge. If we measure
that flow only between the middle and bottom plate, it will be linear with the middle plate’s motion. (At
the end of the chapter, there is an explanation of the mathematics of capacitors.) This is exactly how the
accelerometer works: The tongue from the H-shaped piece corresponds to the middle plate, and the two
stationary fingers are the fixed plates of the differential capacitor. At rest, no current flows. But if we
move the accelerometer, the tongue will jiggle, creating a current proportional to the acceleration. Now
we have the perfect device to put in our phone and tell us which way is up, but that is only half the battle.
The next problem lies in making something so small but so complex.
How to Make an Accelerometer
It would seem nearly impossible to make such an intricate device as the tiny smartphone
accelerometer; at only 500 microns across, no tiny mechanical tools could craft such a thing. Instead,
engineers use some unique chemical properties of silicon to etch the accelerometer’s fingers and H-
shaped section. The method is known as MEMS, or micro electro-mechanical systems. To get an idea of
how they do this, let me show you how to make a single cantilevered beam out of a solid chunk of silicon.
A cantilevered beam is one that is anchored at one end and sticking out over a hole; for instance, a diving
board is a cantilevered beam. The moving section of the smartphone accelerometer is just a complex
cantilevered beam. Let’s start by etching a hole in a piece of silicon.
Empirically, engineers noticed that if they poured potassium hydroxide (KOH) on a particular surface
of crystalline silicon, it would eat away at the silicon until it formed a pyramid-shaped hole.
To make a pyramidal hole in silicon, engineers first cover all but a small square of the (100) plane --
that’s what we meant by a “particular surface” -- with a “mask” of silicon nitride (Si 3N4). The mask is
impervious to the KOH, so the KOH etchant will now only etch within the square shape cordoned off by
the mask until it’s washed away.
This directional etching occurs because of the unique crystal structure of silicon.
The three main steps (from left to right) to make a pyramidal hole in a piece of crystalline silicon. Each step is shown from
three angles (top to bottom): perspective view, top view, and side view of each step.
In a chunk of silicon, each silicon atom has four closest silicon neighbors surrounding it. The four
neighbors create a tetrahedral region around the silicon at the center. Millions of these “units” make up a
slab of silicon. They’re packed together in an open structure that is not the same in every direction. For
example, the number of atoms along the three directions shown above -- labeled [100], [010], and [001] -
- differ in the number of atoms in the planes perpendicular to each direction. The atoms are packed more
tightly along the [111] direction than in the [100] direction. This means that the KOH dissolves both the
(110) surface and the (111) surface, but it chews through the latter much faster. This is why it makes a
pyramidal hole.
Crystal structure of silicon. Every atom in a silicon crystal is identical. Each is bonded to four other silicon atoms. The
darkened atoms in this figure show most clearly that each atom has four nearest neighbors. Note that there are more atoms in
the plane perpendicular to the [111] direction than there are in planes perpendicular to the [100] direction.
To make a cantilevered beam, engineers mask all the surface except for a U-shaped section. At first the
KOH will cut two inverse pyramids side by side.
As the etching continues, the KOH begins to dissolve the silicon between these holes. If we wash KOH
away at just the right point before it dissolves the silicon underneath the mask, it will leave a small
cantilever beam hanging over a hole with a square bottom.
The three main steps (left to right) to make a cantilevered beam in a piece of crystalline silicon. Each step is shown from
three angles (top to bottom): perspective view, top view, and side view.
Although engineers use these principles to produce smartphone accelerometers, they use much more
complicated masks and multiple steps. As you can picture, first a machine would apply a mask, then etch
the silicon, change the mask, and etch some more; the process is repeated until they create an intricate
structure. While complex, a key point is that the whole process can be automated. Engineers now make all
sorts of amazing things at this tiny scale: micro-engines with gears that rotate 300,000 times a minute;
nozzles in ink-jet printers, and, my favorite, micro-mirrors that redirect light in optical fibers.
In Depth: The Mathematics of Capacitors
Simple algebra reveals why the capacitance of a two-plate capacitor changes non-linearly with a change in the plate separation and why a
differential capacitor changes linearly.
The capacitance between two plates, which is just the amount of charge that it can hold, depends on three factors:
1. the distance between the plates,
2. the surface area of the plates,
3. the electrical properties of the material between the plates, known as the dielectric.
The dielectric constant (ε) is the ratio between the amount of electrical energy stored in a material by an applied voltage to the energy
stored if the space between the plates were filled with a vacuum. Often this space is instead filled with a common mineral known as mica,
which has a dielectric constant of around 5; frequently the dielectric between the plates is just air (ε = 1). We define capacitance as
Where ε is the dielectric constant, A is the area of the plates, and d is the distance between the plates.
If we move the plates closer together by a small amount, say Δ, then the new capacitance is
For very small motions, we can use a Taylor Series to approximate this:
The Taylor Series is a useful way to simplify some mathematical formulas under certain conditions. For example,
can be approximated by
if y is much smaller than x. You can test for yourself how much smaller y must be than x for this to be true. Here, we are assuming that the
change in the distance between the two plates will be much smaller than the original distance, so we approximate with a Taylor Series.
That squared term causes a non-linear response, which is most easily seen in the plot below.
This plot shows the response of both a two-plate and a differential capacitor. The y-axis is the fractional change in the separation
of the plates -- here they are getting closer together -- and the x-axis shows the change in capacitance. Note how the two-plate
capacitor deviates from a linear response, while the differential capacitor has a linear response in the same range.
To overcome the non-linear problem, engineers use a differential capacitor. As described in the chapter on the accelerometer, this
capacitor has three plates: one in the center, one above a distance d, and one below a distance d. The capacitance of each plate is then:
and
Of course if the plate is perfectly centered, the difference between these two capacitances would be zero, but it will change if we move the
middle plate down by Δ. That is, as the center plate gets further from the top plate by Δ and closer to the bottom plate by Δ. This difference
becomes:
Using that Taylor Series expansion for each capacitance, we get:
The two squared terms cancel each other out, leaving us with:
As you can see in the plot above, the capacitance now changes linearly with plate separation.
Cesium
[seez-i-uhm] | Symbol: Cs | Atomic number: 55 | Atomic weight: 132.9054 amu
In 1861, Robert Bunsen was on a search for a new element. Having already invented
his famous burner, Bunsen was working with co-discoverer Gustav Kirchhoff (famous
today for his eponymous laws of electricity) to extract a new element from mineral water.
Boiling down 30,000 liters of water, they extracted from it everything they already knew
about -- lithium, sodium, potassium, magnesium, and strontium. What remained was a
thick “liquor” of some unknown substance, which Bunsen placed in the flame of his
burner. Two brilliant blue lines appeared in the flame when viewed through a prism,
indicating a new element. He named it cesium, from the Latin caesius, meaning blue-gray.
The defining characteristic of this soft, silvery metal is its low melting point of 29 oC.
Place some in your hand and it’ll melt, but it is so reactive that it will also burn your
palm: The moisture in your skin will decomposed into hydrogen and oxygen.
Cesium has two traits important in making the first atomic clock, as described in this
chapter. First, its low melting point makes it easy to create a gas. Second, the element has
only one stable isotope (133Cs), so all cesium atoms are identical, allowing it to be used as
a standard in timekeeping.
How an Atomic Clock Works
INSIDE A BUNKER IN Boulder, Colorado, an atomic clock -- blandly named NIST-F1 -- measures time
so accurately that it will neither gain nor lose a second over the next 100 million years. It sends its time
signals to an elegant eighteenth century building, the Pavillon de Breteuil on the outskirts of Paris, which
houses the Bureau International des Poids et Mesures , (International Bureau of Weights and
Measures) keeper of the SI (Système International d'Unités). There, an international team of clock
watchers combine the results of almost 260 clocks to create “Coordinated Universal Time” ( UTC), the
world standard for civil timekeeping.
Our high-tech world depends on accurate time in nearly every respect, but perhaps the most striking
example is the global positioning system (GPS), which requires time be measured to an accuracy of at least
one nanosecond to operate correctly. To achieve that accuracy, every GPS satellite orbiting the earth
contains an atomic clock that marks time with incredible accuracy.
To make a clock extraordinarily accurate requires a standard that will remain the same over long
periods of time. For the atomic clock, it is easy to specify that standard: The energy difference between
the ground states of a cesium atom -- a difference of 3.8 electron volts, or about six-sextillionths of the
energy used when a 100-watt incandescent bulb burns for one second. This energy difference is the same
for every atom of cesium kept at the same conditions of pressure and temperature. The hard part is
engineering a device that uses this standard to make a reliable, useful clock. To understand the
engineering principles used in designing an atomic clock, we need first to understand how to keep time by
resonance.
Basics of Modern Timekeeping
In the past, time was measured using physical actions, such as the flow of water in elaborate sixth
century BC Chinese clocks, or the hourglass’s flowing sand. In our modern age, we measure time by
resonant control. (See figure for a definition of resonance.)
The most familiar example of using resonance for timekeeping is the pendulum of a grandfather clock:
It swings back and forth with a regular period, typically two seconds for the pendulum to complete one
arc. (The period is the time it takes for the pendulum to swing completely in one direction and then back
again to its original starting point.) Based on this period, the manufacturer designs the gears in the clock to
translate this periodic motion, or resonance, into a movement of the hands that marks time in minutes. If
there were no friction or drag, the pendulum would swing forever, keeping perfect time; but, of course,
the arc slowly decays, that is, it gradually stops swinging. On each swing, the pendulum loses a little
energy to friction and does not return to as high a point as it did before. If you wanted a grandfather clock
to keep better time, you could tap it every time the pendulum’s arc decreased just a little bit to put it back
on track.
Resonance is the tendency of a system to oscillate with larger amplitudes at some frequencies than at others; these points
are called a system’s resonant frequencies. All resonators, if moved from rest, transform their stored energy back and forth
from potential to kinetic at a rate depending on the mass and stiffness of the spring or pendulum, or equivalent electrical
properties. At each oscillation, that is, the change from potential to kinetic energy, the resonators lose a small portion of their
energy to internal friction and so eventually decay. There are many ways to create resonance, all of which can be used to tell
time. Shown in this figure are a) an electric circuit with an inductor and capacitor; b) a pendulum driven by gravity; c) a quartz
resonator; d) a tuning fork made of steel or quartz; and e) a hairspring with a balance as used in early twentieth century watches.
In essence, this is precisely how an atomic clock works. It also has a resonator keeping fairly good,
although not perfect, time. Taking the place of the pendulum inside the clock is a piece of quartz, and the
“tap” comes from an electronic signal guided by a device that uses cesium atoms to detect when the
resonator’s period has decayed a tiny bit. Picture the “atomic” part like cruise control. In a car, you use
cruise control to set a constant speed, and if the car speeds up or slows down on a hill, the control system
responds automatically to keep the car’s speed at its set point. In an atomic clock, the speed is the period
of the resonator. If it slows down, a circuit tells it to speed up a bit and vice versa. In engineering
parlance, this is a type of feedback control.
As you might guess, creating the “cruise control” from a cesium atom is pretty complex. Let’s break it
down by looking at the resonator that lies at the heart of an atomic clock.
Quartz Resonators
A pendulum is only one of many ways to create resonance for measuring time. A pendulum system
would be far too inaccurate for an atomic clock, so engineers use a piece of quartz, typically just a solid,
rectangular slab a few millimeters in length and width and fractions of a millimeter in thickness.
It doesn’t seem obvious that it would work as an oscillator, at least not as obvious as a pendulum, yet
the principles are the same. For a moment, instead of thinking about a slab of quartz, think of a rectangular
chunk of Jell-O. If you tap it, it vibrates back and forth. It doesn’t last for long, but it does have a periodic
motion or resonance, and just like the pendulum, its motion stops eventually. A piece of quartz will
resonate in the same way.
A piece of quartz can oscillate (vibrate) in many different ways, each of which has a different frequency depending on the
thickness and dimensions of the crystal. Shown above are some of its main modes of vibration. Most important for the atomic
clock is the thickness shear vibration mode (bottom right in the figure). The most precise crystals vibrate in the fifth overtone of
this mode at frequencies of 5 MHz or 2.5 MHz.
Tap it in the right way (see next figure), and it will oscillate at five million times per second, or 5
MHz. Unlike Jell-O, we don’t tap the quartz to have it oscillate. To start the oscillation of the quartz to
measure its period, engineers use the piezoelectric properties of quartz. The word piezoelectricity comes
from the Greek verb piezein, which means to squeeze or press.
To start the quartz crystal oscillating, we “tap” it with a jolt of electricity (a small voltage across the
quartz) and then measure the vibration from the current it produces. That tap will keep the quartz
resonator going for a long time; for example, its oscillations will last roughly 1,000 times longer than the
swing of a pendulum of a grandfather clock.
A piezoelectric material can convert mechanical motion into electricity and vice versa. For example, if we were to attach two
foil electrodes to a piece of quartz and then strike it with a hammer, the crystal would generate a current. And if, in turn, we
apply a voltage across those electrodes, the crystal will deform.
It might seem the quartz oscillator solves the problem of creating precise time. Quartz is ideal for
clocks because of its outstanding physical hardness as well as its mechanical and chemical stability. A
quartz resonator provides precision only to about 1 second in 300 years for a short period of time
(although the best resonator can achieve 1 second in 3,000 years), but if we combine it with feedback
from an atomic standard, we can make a clock with astounding accuracy.
Cesium-based Atomic Clocks
Recall that our idea was to take an oscillator and, just as the period of its motion begins to decay, give
it a “tap” to restore its oscillations. The “atomic” part of an atomic clock uses cesium to create a way to
determine when the oscillations of a quartz crystal have decayed too much and the crystal needs to be
“tapped.” We do this by measuring a particular property of cesium.
The atoms in pure cesium exist mostly in two slightly different forms: A low energy form and one of
just slightly higher energy. For an atomic clock, these two variations, usually called states, have two
properties critical in making a clock:
1. They can be separated by a magnet because they differ in their magnetic properties.
2. The lower energy atoms can be converted to the higher energy ones if we bombard cesium with
radiation of exactly the right value as characterized by the energy’s wavelength.
In that word “exactly” lies the heart of the atomic clock’s great accuracy. Engineers tie the frequency
of the quartz resonator to the wavelength of the radiation bombarding the cesium. That is, if the
resonator’s frequency changes, then the radiation changes wavelength and no longer converts the lower
energy to the higher energy atoms. This means that if we can detect whether higher energy cesium ions are
being converted from the lower energy ones then we have a feedback mechanism for keeping the quartz
resonator’s frequency accurate. Here’s how it’s done.
In an oven, we heat cesium chloride to create a gaseous stream of cesium ions. The stream contains
both lower- and higher-energy-state ions, so we first pass it through a magnet to separate the two types of
ions. The higher-energy ions are discarded, and the lower-energy ions are allowed to pass into a
chamber. Inside that chamber, we bombard the ions with energy equal to the energy difference between
the two lowest ground states of cesium, which converts the lower- to the higher-energy ions. As these
gaseous ions leave the chamber, they pass through another magnet that directs high-energy ions toward a
detector, discarding any lower energy ones. The detector converts the arriving ions into a current. The key
to understanding how this creates an extraordinarily accurate clock is seeing what happens if the energy
isn’t tuned to just the right wavelength to make the atoms change.
A schematic of the first atomic clock made in 1955. An oven creates a gaseous stream of cesium. This gas contains a
mixture of cesium in its lowest and next-to-lowest energy states. The stream flows through a set of magnets that remove the
high-energy atoms. The lower-energy atoms flow into a chamber where they are bombarded by microwave radiation whose
frequency is tied to the frequency of the quartz oscillator. If that radiation is of the right frequency (or, equivalently, wavelength),
all of the cesium atoms in the lower energy state are converted to higher energy. As they exit the chamber, a second set of
magnets directs these higher-energy atoms to a detector. The signal from this detector provides feedback to the quartz
oscillator: Should the frequency of the oscillator drift, the signal from the detector decreases, indicating the quartz oscillator
needs to readjust its frequency (see the next figure for the signal from the detector). In this way, the quartz oscillator can keep
time such that it loses only a second or so in a million years.
The trick here is to tie the current from the detector to the quartz oscillator. When the oscillation’s
period decays, the energy bombarding the cesium ions in the chamber changes, and fewer high-energy
ions exit the chamber, so the current decreases or stops.
A detector at the end of the chamber measures the number of high-energy cesium ions produced. It converts the number
into a current. If we plot that current versus the oscillator frequency, we can see a clear current spike at the optimal frequency.
When the electronics of the atomic clock detect this current decreasing, they send a voltage to the quartz resonator and, via the
piezoelectric effect, restore its oscillation frequency.
The decrease in current tells the electronics to “tap” the oscillator and correct the period of
oscillation. The electronics create this “tap,” of course, by applying the proper voltage that, via the
piezoelectric effect, deforms the quartz and restores its oscillation frequency.
Uses of Accurate Time
It’s very nice to have clocks that keep such extraordinary time, but of what use are they to engineers?
Well, an incredibly important use of atomic clocks is our global positioning system, used in almost every
facet of life today. The obvious uses of GPS include guiding commercial planes, helping us find a location,
and keeping track of a fleet of cargo trucks. However, GPS reaches even deeper into our world, striking at
the core of our society: food production. Farmers use GPS extensively for precision or site-specific
farming. This technology allows farmers to apply pesticides, herbicides, and fertilizers precisely, which
reduces expenses and increases yield. Before GPS, a 4,000-acre farm needed eight or nine tractors; today
the same farm needs just three or four.
Throttling a GPS Device
All technological advances come with promise and peril. A GPS navigation unit makes life easier, whether it’s a trip for pleasure
or tracking a trucking fleet so its operations can be optimized. But in the great, even astonishing, precision of GPS lies the potential
to guide a weapon exactly to its target. In the United States, the Department of Commerce requires all exportable GPS devices to shut
down when they travel faster than 1,000 knots or when their altitude is above 18,000 meters. This restriction prevents GPS devices
from being used to guide missiles to targets.
Where d1, d2 and d3 are the distances between the receiver and each satellite, which are obtained using
the speed of light as described above; and xi, yi, and zi are the coordinates of the satellites, which are
embedded in their signals. Because all three satellites have synchronous time from their atomic clocks,
we only have three unknowns: the positions of the receivers xr, yr, and zr. Note that if the satellites
weren’t synchronized we’d have three more variables, an insolvable problem! These three equations
aren’t enough: Every one of the distances is wrong because the clock in the receiver is not accurate.
In principle, the receiver can use simultaneous signals from three satellites to determine its position, but as noted in the
text it needs a fourth satellite to correct the receiver’s clock. Using information from the first satellite’s signal -- the time it was
sent and its position -- the receiver can determine that it lies on the surface of a sphere of radius d 1 . From the second satellite,
it creates another sphere of radius d 2 -- the distance the receiver is from the satellite. The receiver must be where those two
spheres intersect, as shown by the circle represented by the shaded area. The location must lie on that circle. The distance from
the third satellite creates another sphere, whose interaction with the first two spheres narrows the receiver location to two
points: One can be rejected immediately because it is far from the earth’s surface, the other is the correct location of the
receiver.
We’ve assumed above that the receiver is synchronized with the satellites, which it would be if it had
an atomic clock built into it, but, as noted above, to keep the receiver cheap and compact it has only a low
accuracy clock. So, we need to correct the distances by the error in the receiver’s clock. If the receiver
differed from satellites by Δt then the distance error is cΔt where c is the speed of light. Our three
equations then become:
We now have three equations and four variables: xr, yr and zr, and now Δt. We need a fourth equation,
which we get from a fourth satellite. Just as with the first three satellites, the receiver calculates its
position from this fourth satellite:
We now have four unknowns and four equations. These four equations are highly non-linear and cannot
be easily solved in the usual way by substituting. Instead, engineers use a procedure called Newton-
Raphson iteration.
To implement this iteration, they expand each equation into an infinitely long polynomial based on the
trial set of values for the position. These four linearized equations can be solved simultaneously to
determine whether this guess was right; if not, we substitute the values generated and repeat. If the
receiver starts with a good guess of where it is -- usually its last location -- then this can often converge
in one iteration. If you instead used the GPS receiver, turned it off, and then flew half way around the
world, it would take a bit of time for the receiver to locate itself.
Primer: Nuclear Structure
IN THE CHAPTER THAT FOLLOWS, we discuss the engineering concepts behind enriching uranium for
use as fuel or in a nuclear bomb, but to set up our understanding of this process, we must have an
understanding of the structure of atoms in general.
All atoms consist of a nucleus composed of protons and neutrons surrounded by a cloud of negatively
charged electrons. The protons are positively charged, while the neutrons have no charge, but they are
each 2,000 times the size of an electron. Thus, the mass of an atom is pretty much just the size of its
nucleus. Several forms of many atoms occur naturally, and we call these different forms isotopes.
Isotopes differ from one another in the number of neutrons each isotope contains; for instance, while the
most common naturally-occurring form of the element carbon has 6 neutrons, the isotope of carbon that
has 6 protons and 8 neutrons in its nucleus (used for carbon dating) is written as 14C, where the
superscript lists the number of protons and neutrons together. Uranium, the element of discussion in this
chapter, occurs in nature as mostly isotope 238U (~99%), with a small amount of 235U (0.72%).
The nucleus is filled with positively-charged protons, which repel each other. The force that keeps the
nucleus from flying apart is known as the strong nuclear force. This is one of four fundamental forces of
nature (the other three being the weak nuclear force, electromagnetism, and gravity) and is also the
strongest. It is incredibly powerful at short distances, but at distances more than a few angstroms (1 x
10-10 m) long, the force drops to zero. The balance between these two opposing forces -- protons
repelling each other and the strong nuclear force keeping them together -- helps to explain why some
elements can undergo fission.
This graph shows the amount of binding energy per nucleon. (A nucleon is either a proton or a neutron.) The x-axis shows
the total number of nucleons, which equals the atomic mass. Thus, the elements appear from lightest to heaviest from left to
right on the x-axis. The greater the binding energy, the more stable the nucleus. The peak occurs at ~60, which corresponds to
iron, the most stable element in the universe. Elements to the left of iron have a tendency to combine (fusion) to become more
stable. For example, in the sun, hydrogen, the lightest element, undergoes fusion to become helium, the second lightest
element. Elements to the right of iron can become more stable by splitting apart (fission).
Fission
In undergoing fission, a nucleus splits apart, releasing the energy binding it together. Elements with
atomic numbers larger than iron are less stable than those with fewer protons. As the atomic number
increases, the number of protons being packed together increases, yet geometry indicates that they must
become further apart as their number increases. This means that the strong interaction is diminished, yet
every pair of protons in the nucleus stills feels an electrical repulsion. Thus, the force that holds the
nucleus together weakens, and the force driving it apart increases. This makes atoms of high atomic
number susceptible to fission. Nuclei are typically stable, so fission requires adding energy to the
nucleus.
Typically, we break up a nucleus by bombarding it with neutrons. That sounds exotic, but it’s good
enough to think of the bundle of neutrons and protons at the nucleus’s core as a rack of billiard balls and
the bombarding neutrons as a cue ball. Just as with cue balls, the degree to which the rack breaks up
depends on the speed of the cue ball; the analog in nuclear physics is slow and fast neutrons. Bombarding
235U with slow neutrons causes its nucleus to begin to vibrate violently because the kinetic energy of the
moving neutron is transferred to the particles in the nucleus. The uranium atom then breaks apart:
235U + neutron → 90Sr + fission fragments
+ 2 or 3 neutrons + energy
The fragments can be many things, but typically 235U fissions into Strontium-90 (90Sr). Not all nuclei will
easily undergo fission. Most elements are stable so that when bombarded with neutrons they simply
absorb the neutron and decay later, or they require very high energy (fast) neutrons. To extend the analogy
with billiard balls a bit, imagine the rack of balls being “sticky” so they require a faster (higher kinetic
energy) cue ball to break them apart. To extract nuclear energy in a practical way requires elements that
are reasonably stable so they can be stored without appreciable decay but are capable of undergoing
fission with neutrons of all energies. Only four elements meet these criteria: 235U, 233U and the plutonium
isotopes 239Pu and 240Pu.
You can visualize fission by picturing the uranium atom as a spherical liquid drop before the neutron hits; after impact, it
vibrates, becoming elliptical and stretching until it looks like a Q-tip. Eventually the thin section breaks and the nucleus splits --
fissions -- into smaller parts, with splinters flying off. For 235 U those splinters are typically two or three neutrons.
destructive power, a bomb maker needs to remove the 238U in natural uranium so it becomes enriched in
235U. For uranium to be used as fuel for nuclear power, it must be enriched to 4% 235U. However,
enriching natural uranium to 90% 235U provides enough of the fissile 235U to power a nuclear bomb.
Luckily this is a difficult engineering problem to solve.
The Power of the First Nuclear Bomb
The first nuclear bomb, which destroyed Hiroshima, contained about 60 kilograms (about 132 pounds) of uranium 235 U, of
which only 600 g (1.33 pounds) -- or about 100 sextillion 235 U atoms -- underwent fission. This was enough, though, to create an
explosion equal to 13 to 18 kilotons of TNT.
where vrms is the root mean square velocity (a type of average velocity), k is Boltzmann’s constant, T is the absolute temperature,
and m is the mass of the particle.
Because the velocity is inversely proportional to the square root of the mass, a lighter particle will travel faster than a heavier
one. This means that in a gas mixture of two particles of different weights, the lighter one will have a higher probability of hitting a
hole in the barrier. It has a higher probability because it travels further on average than does a heavier molecule; this typical
distance traveled, called the mean free path, sets the size of the holes in the barrier.
At atmospheric pressure, the mean free path of a molecule is about one ten-thousandth of a millimeter or one-tenth of a micron.
This is just another way of representing the distance that, on average, a molecule travels before colliding with another molecule that
changes its direction or energy. Thus, to ensure the necessary separation for an atomic bomb, the diameter of the billions of holes in
the barrier must be less than one-tenth the mean free path, otherwise both effusing isotopes would pass through the barrier. The
exact composition and preparation of the barriers used today in separating uranium are secret, but for earlier barriers, engineers
rolled Teflon powder onto nickel gauze to create pores of 0.015 microns.
Since 1993, the number of transistors on an Intel-brand computer CPU has increased from 3,100,000 in 1993 on their first
Pentium chip to 291,000,000 on their 2006 Core 2 Duo. In contrast, the energy density of a lithium-ion battery has increased
over the same time period from 89 to a little over 200 watt hours (Wh)/kg.
Are Electric Cars Really Better for the Environment?
A focal point in the discussion of climate change is the emission of greenhouse gases, especially the contribution of carbon
dioxide to the warming of the atmosphere. A popular “green” choice to alleviate some of the environmental impacts of fossil fuel-
powered cars is the use of electric or hybrid electric cars. However, the case for such cars isn’t as clear as you would expect. The
proper way to assess the impact of an object on the environment is life-cycle analysis; that is, looking at its environmental impact at
every stage of its life, from birth to death.
Several researchers have compared the impact of traditional gasoline-fueled vehicles, electric cars, and hybrids like the Toyota
Prius over their lifetimes. These studies show that while electric cars produce far fewer emissions throughout their performing life,
they take an incredible amount of energy to produce. Wired Magazine reported that the Toyota Prius consumes over 1,000 gallons
of gas before it even drives its first mile! The nickel batteries that power the car require lots of fossil fuel energy to come to life. The
nickel is mined in Canada, travels across the world to Europe for processing, and then on to China for further processing. From
there, it is transported to Japan where the cars are manufactured, and then finally to the United States to be sold. These trips, as well
as the processing taking place at each stop, release emissions that begin to rack up before the car even hits the dealership. Factor
into this equation the small operating life of the battery (typically no more than 100,000 miles), and the electric car doesn’t look as
green as it once did. Ultimately, it’s tough to say whether or not the electric or hybrid car outperforms a new gas car, but it is likely
that a fuel-efficient used car is better right now than either option because the first owner has already paid off the carbon debt of
the car’s manufacture. And with dozens of cars from the 1990’s that top out at over 30 miles per gallon, you have a decent selection
to choose from. The overall lesson here? Like many of today’s problems, there is no perfect solution available yet.
The electrochemical reactions on the left generate no useful electricity from the exchange of an electron between lead and
lead oxide. The net current (flow of electrons) is zero. To make a battery, an engineer needs to arrange things so the electrons
flow through a wire. As shown on the right, the basic components of a battery are: Positive and negative electrodes linked by a
wire outside of the solution, an electrolyte solution that can transport charge between the electrodes using ions, and a separator
that allows the ions to pass but keeps the electrodes separated.
Elements of a Battery
To get useful work from the conversion of chemical energy into electrical energy, we need to arrange
the constituents above in a special way. Specifically, we shape the lead and lead oxide into separate solid
electrodes and keep them separated. If we place these two electrodes into a water/sulfuric acid solution
and connect them with a wire, we find that a current will flow along the wire. The ions H+ and SO42-
carry a charge between the two electrodes, while the electron that needs to be exchanged between the
lead and lead sulfate electrodes is carried through the wire. This is all fine in principle, but it doesn’t
make a practical battery that can be sold.
Electrolytes
We call the sulfuric acid/water solution the electrolyte. An electrolyte is any substance that has free ions that make the substance
electrically conductive. Typically, this is an ionic solution, but electrolytes can be made from molten salts and even solids. For
example, negatively-charged oxygen atoms in solid yttria-stabilized zirconia, a zirconium-oxide based ceramic, can act as charge
carriers.
Typically an explosive contains both oxidizing and reducing sections within the same molecule. Shown on the left is
nitroglycerin used in dynamite, the first practical explosive ever produced that was stronger than black powder. On the right is
trinitrotoluene, or TNT, one of the most commonly used explosives because of its insensitivity to shock and friction.
The structure on the left is unprotected. Without cathodic protection, the iron in the steel pillar can corrode. In salt water, the
elemental iron loses two electrons to become an ion, which dissolves in the water. The two electrons produced are conducted
through the pillar, where the dissolved oxygen and water react to form hydroxide. These can form Fe(OH)2 and the familiar
oxides that appear when iron rusts. The structure on the right is protected. By supplying electrons to the pillar, we can prevent
the iron from dissolving. Typically an offshore rig will have a large metal oxide electrode next to the pillar. A battery drives
electrons from the new electrode to the pillar.
Why Our Buildings Don’t Fall Down (Or, Why Stainless Steel Doesn’t Rust)
Although mainly composed of iron that typically rusts rather easily, stainless steel corrodes very, very slowly. Called the
“Miracle Metal” and the “crowning achievement of metallurgy, ” stainless steel truly makes our modern world possible: Buildings,
ships, cars, and likely even your kitchen sink are made of the stuff. Reflect for a moment on what a wonder it is: Imagine being an
alchemist in the Middle Ages and being told that a material exists that looks like platinum, resists chemical attack like gold, and can
support a quarter million pounds per square inch. You’d think the inventor was the greatest miracle worker of all time! The secret
element in stainless steel is chromium.
Stainless steel contains, by definition, at least 10% chromium by weight. Chromium reacts with water and oxygen to form a thin,
tightly-bound stable protective film (similar to the one found on the Delhi Pillar and Apple’s computers) that covers aluminum. Since
the chromium is mixed through the iron, and not just on the surface, the protective layer will grow anew if the surface is scratched or
if the piece of steel is cut to be shaped into something else.
Anodizing Aluminum
We typically think of aluminum as inert and unlikely to corrode or rust (think of aluminum foil or a
bike frame), yet from a chemical viewpoint, aluminum is highly reactive. Sodium hydroxide dissolved in
water will attack aluminum vigorously; calcium hydroxide gives off hydrogen almost immediately on
contact with the metal. Hydrochloric and other halogen acids attack aggressively; when exposed to the
chloride ions in sea water, aluminum will corrode within a few years. Yet we find it an inert metal in our
technological world because engineers control the corrosion in a way that it forms a substantial protective
layer.
Left exposed to the air, oxygen will attack aluminum’s surface. The resulting aluminum oxide film
keeps the aluminum from corroding further because it seals completely, at the atomic level, with a barrier
film that doesn’t conduct electricity. Recall that for corrosion to occur, electrons must move. The
naturally-formed film is very thin and can be removed easily by scratching the surface. To create
industrial-strength coatings engineers enhance this film electrochemically.
To do this, they make aluminum a negative electrode in an electrolytic cell. Current flows so that it
drives the aluminum to a negative potential by using a second, largely inert, metal as an electrode. This
current causes an aluminum oxide layer to form. At first it forms the thin “barrier” layer similar to what
forms naturally. But then as the electrolysis process continues, the current “pushes” this barrier down
into the aluminum, converting the aluminum above and growing a very porous oxide layer. This new layer
is made from the aluminum: It isn’t a layer being put on top but instead grows into it in a manner that
consumes the aluminum. This is one of the reasons it’s so effective at keeping aluminum from corroding.
These pores give the aluminum a unique characteristic important for a consumer device: the ability to be
colored or “decorated” to use the engineering term. The pores formed on the surface have a honeycomb
pattern.
Anodizing occurs in an electrochemical cell. The piece to be anodized is used as the positive electrode. As electrons are
removed from the positive electrode, a layer of Al2 O3 grows into the aluminum part being anodized.
The center hollow section is typically 300 or 400 angstroms in diameter and perhaps 1,000 angstroms
or so deep. (100 angstroms is a millionth of a centimeter.) Inside these layers, one can place dye of any
color. Once the pores are filled, engineers
A piece of aluminum has a thin protective layer of Al2 O3 that forms naturally on it. (See upper right-hand corner.) When
anodizing takes place, this layer becomes thicker, then eventually forms a porous layer with hexagonal pores. Dye can be sealed
into these pores.
seal the layer by boiling the aluminum in hot water. This closes the pores, locking the color in forever;
you cannot scrape the color off without removing the aluminum, as it is a thick, tough oxide coating.
Without the sealing, only the thin barrier lying adjacent to the metal would exist. Just like the natural
barrier, it isn’t thick enough to provide robust protection, but the sealed porous layer creates an extremely
tough surface with or without dye. Its toughness likely comes from the oxide's relationship to tough
gemstones. Sapphire is Al2O3 -- an aluminum oxide -- with trace amounts of iron and titanium to give it
a blue color; the same Al 2O3 structure with chromium that absorbs yellow-green is also the basis of ruby.
Both materials are very hard: 9.0 on the Mohs scale, which ranges from 1 for soft materials such as talc to
10 for diamond. Apple typically uses a thinner oxide layer resulting from what’s called soft anodizing.
This isn’t as durable as the thick layers used for industrial products.
Anodizing Titanium
Titanium, a metal with many properties similar to aluminum, can also be anodized with beautiful
results. Unlike aluminum, the thickness of titanium’s oxide layer rather than dye determines its color. In
titanium, a much thinner, transparent oxide layer forms than on aluminum, which gives titanium a wide
array of colors due to thin-layer interference.
The oxide layers in titanium appear colored because of thin-layer interference. It’s the same phenomenon that causes the
iridescence in opals, oil slicks, soap bubbles, peacock feathers, and rainbow trout. When light rays strike the titanium surface,
some pass through the transparent film created by anodizing and reflect off the surface, and others reflect from the outer edge of
the film. If these two rays of a particular color combine when they are a half wavelength out of phase -- where the crest of one
meets the troughs of the other wave -- the color observed will be white minus that color, thus appearing as its compliment.
For example, the red in rainbow trout comes from the destructive interference of cyan, whose compliment is, of course, red.
This means that the color of anodized titanium depends on the thickness of the layer.
Primer: Waves
THROUGHOUT THIS BOOK, we often mention waves in one way or another. Here we will cover a few
basic principles of waves so that we can use waves to further our discussion of the electromagnetic
spectrum.
Basic Definitions
Start by imagining a buoy in a shallow part of the ocean. It is, perhaps, 10 meters or so below the
surface and can move up and down on a pole that keeps it from moving left or right. As you would expect,
it will bob up and down with the motion of the ocean. If we were to measure its position on the pole over
a long period of time, we would see the buoy rise to a maximum height, then fall to a minimum height, and
repeat this motion again and again.
We can draw a line between the two extremes that neatly divides the buoy’s position: Half the time it
is above this line, half the time below. The deviations of the buoy from this level are called its amplitude.
We call the time it takes for the buoy to start at its maximum positive amplitude, pass through the
maximum negative amplitude, and then return to the maximum positive amplitude the period T; that is, the
time between two “crests.” In the drawing below, this happens in two-tenths of a second. Often, instead of
talking about the time it takes for these two crests to appear, we’ll talk about how many crests we would
observe in a second. This is called the frequency, f , and is related to the period T by f = 1/T. For the buoy,
the frequency is 5 crests/second.
The motion of a buoy with time in a shallow part of the ocean. The buoy is on a rigid pole so that it can move up and down but
not from side to side. The figure shows the same buoy at seven different times - a, b, c, d, e, f, and g.
Next, imagine a series of these buoys fixed along the ocean bottom. If we take a snapshot at various
times, we can find the wavelength of the wave. We call the distance from one crest to the next in any
snapshot the wavelength, usually designated by the Greek letter λ.
A series of buoys affixed to the ocean floor. Shown top to bottom are snapshots of the buoys at different times. Note that here
the horizontal axis is the distance between buoys, where in the previous drawing it was time. You can construct the graph shown
in that figure by looking at the motion of one of these buoys with time. For example, the positions of the buoy at 0.5 seconds -
- labeled a, b, c, and d -- are shown on the graph on the right.
The wavelength and the period (or frequency) are related. The period is the time it takes a wave to travel
one wavelength. Unsurprisingly, then, these are related by the velocity, v, of the wave: λ = vT. In the ocean,
at a depth of 10 meters with a 1 m wavelength, the velocity of the wave will be 1.25 m/sec.
A wave, of course, transports energy. You can see this by the motion of the buoy: The wave’s energy
slides the buoy up and down the pole. More profoundly, think of how the ocean’s waves erode a beach.
Interaction of Waves
We’ve talked above about a single traveling wave so that we could clearly define its properties, but
the most important aspects of waves occur when they interact. Two or more waves can traverse the same
section of space independently. Since they are independent, we can just add their amplitudes at the points
in space or time where they overlap.
The superposition (combination) of two waves (a and b) to create a new wave (c). To construct the new wave simply add the
amplitudes of the individual waves.
Technologically, the most important interaction of two waves is to form a standing wave; it’s what
heats the food in a microwave oven and what allows for the amplification of light in the cavity of a laser.
It’s best seen in a one-dimensional wave like we’ve discussed above. Picture one wave traveling along a
line from the left to the right side of the page and one wave traveling along the same line but in the
opposite direction. To keep things simple, assume the waves have the same wavelength. Now, if they start
from different distances from the left and right, they will be out of phase and the waves will cancel each
other out.
If, though, they started from spots equidistant from the center (and we’ll explain in a moment how that
happens), they will meet in phase, and thus the crests and troughs of the waves will match up. The result
will be a wave that has points that are always zero, called nodes, and the resulting combined wave will
oscillate between these points. If we imagine a particle sitting on a section of the wave, it would rise and
fall but not move left to right.
In engineered devices like the microwave and laser, a standing wave usually results from a wave
being reflected. For example, this is done in a laser by the mirrored ends of the cavity and in a microwave
oven by the metal walls. Specifically, reflection allows only waves with an integral number of half
wavelengths to form standing waves. Waves of other wavelengths simply die out as the waves with
allowed wavelengths grow larger and larger as their crests reinforce each other.
The bottom of (a) shows waves traveling from the right and left, the top shows their sum. (b) Shows the two waves shifted
relative to each other by 90 o . (c) Shows the two waves shifted 180 o out of phase with each other. They cancel each other so
their sum creates a “wave” of zero amplitude.
Shown above is a complete oscillation of a standing wave. The small graphs show how the total energy of the wave is
distributed between potential energy (U) and kinetic energy (K). A standing wave does not appear to move; instead, the
traveling waves that comprise it superimpose to create a wave with fixed nodes -- spots where the wave’s amplitude is always
zero -- and anti-nodes, which smoothly change from the maximum positive amplitude to the maximum negative amplitude. A
particle sitting atop the anti-nodes would be bounced up and down. It would come to a full stop as the anti-node reaches its most
extreme points - the maximum positive or negative amplitude. At that position, the particle would have its energy stored
completely as potential energy. As the anti-node changes, the particle would have its maximum velocity just as the wave passed
the horizontal axis. Here its energy would be completely kinetic. Compare this figure to behavior of an LC-circuit as described
in the chapter on microwave ovens.
Electromagnetic Waves
In this short primer, we’ve only shown a transverse traveling wave. The wave is transverse because
the wave motion is perpendicular to the direction of travel. This type of wave motion is easier to
visualize than longitudinal waves, where the motion is in the direction the wave travels. You can make
both types with a Slinky: Fix it at one end and shake it side to side, and you have a transverse wave; in
contrast, pulsing the fixed end in the direction of travel creates a longitudinal compression wave where
the curls of the Slinky push against each other in order to facilitate energy transfer.
Waves come in many shapes and sizes, but usually engineers talk about sinusoidal waves, as we’ve
shown above, that vary smoothly with time and distance as they pass by us. Any wave can be composed
of the sum of many sinusoidal waves; thus an understanding of simple sinusoidal waves is all that’s
required to understand any wave.
Waves come in many different forms; for instance, you produce waves whenever you disturb water,
pluck a string, or even make a sound! However, the waves we focus on in this book are electromagnetic
waves. Electromagnetic waves differ from mechanical waves -- for example, waves in water -- in two
key ways. First, they need no physical medium to travel through, so they can propagate through a vacuum.
Second, they always consist of both an electric wave and a magnetic wave together.
Any wave can be created by adding together an infinite number of sinusoidal waves. Here the five waves shown individually
at the bottom of the figure combine to create the waveform at the top.
An electromagnetic wave appears when either a magnetic field or an electric field oscillates.
Regardless of how it’s generated, once the wave leaves its source, it is a combination of electric and
magnetic fields in a well-defined relationship: In a propagating electromagnetic wave, one doesn’t exist
without the other. Electromagnetic waves of different frequencies make up the electromagnetic spectrum,
which is the full spectrum of waves, categorized from smallest wavelength to highest. Although an
electromagnetic wave may sound like a complex thing, we are all familiar with one electromagnetic
wave: visible light!
The electromagnetic spectrum contains all possible frequencies of electromagnetic radiation. It extends from low
frequencies used for radio communication to gamma radiation at the short wavelength (high frequency) end; thus it spans
wavelengths from thousands of kilometers to a fraction of the size of an atom.
Tungsten, Thorium, & Copper
[tung-sten] | Symbol: W | Atomic number: 74 | Atomic weight: 183.85 amu
[thor-i-uhm] | Symbol: Th | Atomic number: 90 | Atomic weight: 232.0381 amu
[kop-er] | Symbol: Cu | Atomic number: 29 | Atomic weight: 63.546 amu
Tungsten began life as a nuisance: German miners nicknamed it wolfram (wolf-dirt)
because it interfered with the smelting of the tin they desired. The residue of that name is
still evident in tungsten’s chemical symbol “W”. Later in life, chemists rechristened it tung
sten, Swedish for heavy stone, an appropriate name for a material that rose to prominence
as an element of war. Add tungsten to iron, and one makes an amazingly strong steel that
was first used in creating cannons and guns for World War I. In fact, today’s armament
makers still use tungsten to make armor-piercing bullets. Aside from iron, tungsten is the
metal most used industrially because of a single distinguishing characteristic: No other
metal has a higher melting point, an astonishing 3,422 oC. Thus, it can be heated to burn a
bright white; small wonder it forms the filament of an incandescent light bulb. This same
property makes it ideal for the vacuum tube powering a microwave oven.
However, tungsten alone doesn’t produce enough electrons for its use in the
microwave, so it’s impregnated with thorium. Named for the Scandinavian god of
thunder, Thor, this silvery-white metal has a low melting point but burns brightly since it
is a good emitter of electrons. In the microwave oven’s vacuum tube, thorium is
combined with tungsten to get the best of both metals: a high melting point filament rich
with electrons. The heated tungsten-thorium filament creates great heat, so it requires a
material that can both conduct electricity and heat well. For this purpose, nearly nothing
surpasses copper.
Copper is the most versatile metal on earth because it uniquely unites five properties --
other metals have some, but none have them all -- to make it perhaps the most important
metal. It doesn’t react with water, can be bent to any shape, plays friendly with other
metals to form alloys, and conducts heat and electricity well. One would think the name
of this powerful metal would come from some marvelous Greek or Roman god, but
instead it was named for the island of Cyprus (cyprium in Latin) where it was first found.
How a Microwave Oven Works
PICTURE A VACUUM TUBE . Likely, you are thinking of an old‐fashioned radio. Perhaps you imagine a
huge wooden cabinet from the 1930s, filled with glowing glass tubes, with a vision of the inevitable
replacement of these tubes by tiny transistors and microchips. Yet it’s too soon to relegate vacuum tubes
to the museum. Nearly every American home uses one every day in a microwave oven.
What is a Vacuum Tube?
Before the age of the solid‐state transistor and the microchip, electronic devices used vacuum tubes, also called thermionic
valves. In any electronic circuit, an engineer wants to control the flow of current to do something; for example, to amplify a signal or
create an electromagnetic signal. (Contrast this with an electrical circuit, where current simply flows through a wire.) Vacuum tubes
gave rise to the first wave of electronic technology, especially radio and television broadcasting. Today, of course, solid‐state
devices like transistors have largely replaced vacuum tubes. The solid‐state devices last longer and are smaller, more energy
efficient, and cheaper to make. An exception, as noted in this chapter, is when used in applications where greater power must be
generated.
The vacuum tube, called a magnetron, produces electromagnetic radiation that heats food. Although the
solid state revolution took over and miniaturized most electronic components, the microchip cannot easily
replace tubes for producing power (recall that power is the rate at which energy is transferred). The
signals transmitted from satellites require vacuum tubes, otherwise they would be so weak they could not
reach the earth. The broadcast of AM and FM radio uses vacuum tubes. Radar used for weather forecasting
or military surveillance must use vacuum tubes, as do the navigation systems used by commercial aircraft.
All these applications require a great deal of power, so it’s no surprise that the microwave oven, which
needs power to heat food, also uses a vacuum tube.
In a vacuum tube, an external power source heats a metal cathode so that electrons “boil” off. (This is what gives vacuum
tubes their glow and also causes older electronics to become warm.) The positively‐charged anode, called the plate, captures
these electrons. Between them is a wire mesh called a grid. By applying a positive or negative potential to this, relative to the
cathode, we can control the flow of electrons. A glass enclosure maintains a vacuum around these elements; without a vacuum,
the electrons would not reach the anode. As shown above, it is wired to be an amplifier. The varying potential on the grid (the
signal) is amplified by the electrons flowing from cathode to anode.
As the radiation reflects back and forth from the metal walls of the oven, a standing wave will form
inside the cavity. This standing wave causes hot and cold spots to form inside the oven, as the waves
constructively and destructively interfere with each other, resulting in spots that have greater energy than
others. Although the three‐dimensional pattern of waves is difficult to predict, the principle can be seen
by looking at the waves in a single dimension.
The wavelength of the microwaves used in a kitchen is 12.2 cm. This means that in a cavity 38.1 cm
long (a typical size for a microwave oven), three peaks and three valleys—called anti‐nodes—of the
wave will fit inside. The peaks and valleys represent the greatest energy of the wave, while at the nodes
where it crosses the x‐axis there is zero energy. In our simple one‐dimensional model, these occur at half‐
wavelengths—6.1 cm for a microwave oven. Nodes correspond to “cold” spots; that is, regions where
there is no heating. You can see this in your own microwave oven. Remove the turntable and insert a glass
tray covered with a thin layer of mozzarella cheese and zap it for fifteen seconds. When you remove it,
you’ll see that some of the cheese has melted but some of it has not. This shows the hot and cold spots
corresponding to the anti‐nodes and nodes in the standing wave.
A microwave oven contains three main components: 1. a vacuum tube, called a magnetron, that generates high‐frequency
radio waves; 2. a waveguide that directs that radiation to the middle of the chamber, where the food is placed; and 3. a chamber
to hold the food and contain the microwaves. The chamber has metal sides that the microwave radiation cannot penetrate.
Only standing waves can exist inside a microwave oven, that is, waves whose half‐wavelength is an integral number of
lengths of the cavity. For the wave shown, three half‐wavelengths fit inside the oven’s chamber. The metal walls reflect the
wave; at the walls, the amplitude of the wave is zero.
Ampere observed that two parallel wires with current flowing in opposite directions attracted each other; this meant there
was a magnetic force between the wires. These two experiments showed that a change in magnetic field induces a current, and
in turn the flow of a current generates a magnetic field.
These two experimental observations were unified by Scottish physicist James Clerk Maxwell in one
of the greatest and most technologically important scientific theories of all time. Maxwell gave a unified
description of how electric and magnetic phenomena unite to form electromagnetic waves that can travel
through space, even through a vacuum. Most of our understanding of electricity and magnetism comes
directly from Maxwell’s work.
Why a Microwave Oven Heats with 2,450 MHz Radiation
In principle, many different energies of radiation could be used to heat in a microwave oven. The reason we use 2,450 megahertz
radiation in our ovens is a mix of scientific, social, economic, and political reasons—something true of almost any engineered object.
In fact, a hallmark of something engineered is that it fits into our world—social, political, economic, scientific—unlike a scientific
fact, which is either true or false. The only way to answer why a mass manufactured object operates like it does is to trace its history.
In the aftermath of World War II, all sorts of devices appeared that create electromagnetic radio—television, communication
links, two‐way radio, and radar. In the United States, the Federal Communications Commission regulated what devices could use
what part of the electromagnetic spectrum. For microwave heating, which falls under the category of ISM applications (industrial,
scientific, and medical), the FCC allowed two frequencies: the S‐band (2,450 MHz) and the L‐band/UHF (915 MHz). Each had a
major corporate proponent: Raytheon, which made radar during the war, argued for 2,450 MHz, while General Electric wanted 915
MHz. Although Raytheon had made microwave ovens since the late 1940s, the sales of such ovens sputtered. In the 1960s, it wasn’t
clear who would make the first truly successful microwave oven. General Electric’s ovens in the 1960s used 915 MHz radiation,
while Raytheon stuck with 2,450 MHz. Which is better? It depends on what you are trying to do. Raytheon argued that 2,450 worked
better for cooking small loads—like a hot dog—while GE argued that for large load (like a roast) penetration of 915 MHz was better
and that it created less thermal runaway when defrosting. 915 MHz radiation penetrates food five times deeper than 2,450 MHz
radiation, which means it can cook a large piece of food with minimal burning on the outside. But, as one Raytheon engineer said,
915 MHz radiation “won’t cook bacon worth a damn.” Also, the wavelength of 915 MHz radiation is nearly three times longer than
for 2,450 MHz (32.8 cm versus 12.2 cm). This means that a typical 915 MHz oven would be larger than a 2,450 MHz oven; a 915
MHz oven might have to be four feet wide to be sure even heating occurred.
In a way, social changes dictated the choice between these alternatives. With the rise in the 1970s of households where both
parents worked, the microwave oven became a “re‐heater” of small amounts of food, and thus a home kitchen appliance, rather than
a large oven used commercially in a restaurant. These uses, then, made the 2,450 MHz radiation the standard for microwave ovens.
In showing a continuum of electromagnetic radiation, Maxwell revealed something truly revolutionary:
Electromagnetic waves could travel without a medium; that is, they could travel in a vacuum. Prior to his
work, it seemed that a wave needed a medium such as a gas or a liquid to propagate in. For example, the
waves in an ocean require the medium of water, otherwise that type of mechanical wave has no meaning.
Maxwell’s work opened up the notion that electromagnetic waves could fly through space, which led to
the use of electromagnetic waves for radios, mobile phones, television, and so on. Further, Maxwell’s
laws reveal to engineers exactly how to produce and detect electromagnetic waves.
An electromagnetic wave appears when either a magnetic field or an electric field oscillates.
Regardless of how it’s generated, once the wave leaves its source, it is a combination of electric and
magnetic fields in a well‐defined relationship: In a propagating electromagnetic wave, one doesn’t exist
without the other. So, to create radio frequency (RF) radiation, we need to get an electric or magnetic field
to oscillate at 3,000 times per second or greater. In principle, you could do this with a wire and a battery.
Hook the battery up to the wire, let the current flow, and then reverse the flow by exchanging the wires on
the battery’s terminal. The problem is that radio waves oscillate at least 3,000 times or so per second; no
human hand could move that fast! As you might expect, engineers design devices to make this switching
happen without human intervention. For an engineer, the key in any device is to produce electromagnetic
waves of the right wavelength or frequency.
Maxwell’s work showed that a continuum exists among all types of electromagnetic radiation and that all electromagnetic
radiation has a common mathematical description and origin. Typically engineers divide the continuum into segments based on
the radiation’s wavelength because the useful behavior of the waves depends on its wavelength. The most important divisions
are radio, microwave, infrared, the visible region, ultraviolet, x-rays, and gamma rays. Higher frequencies have shorter
wavelengths, and lower frequencies have longer wavelengths.
If the anode has no cavities carved in it, an electron will follow a circular path that in turn circles the cathode.
Third, the magnetron is operated with the magnetic field set such that the electrons traveling from the
cathode to the anode just brush the edges of the anode. Cut into the anode are circular or triangular
cavities with only a narrow opening to the interaction space. As the orbiting electrons pass by these
openings, they create oscillations within the cavities. Think of it like this: When you blow over the top of
an empty bottle, it makes a deep sound because you are setting up or creating a standing wave or
oscillations in the bottle. The electrons do the same thing as they sweep past the slots, although here the
oscillations are in the electromagnetic region of the spectrum, whereas when you blow over a bottle, you
are making sound waves.
How Heating Tungsten Produces Electrons
To understand the materials of construction for the magnetron or why it is built out of tungsten and thorium, let’s examine a few
critical properties of these elements. Energy supplied to a metal can liberate electrons from the solid’s surface into the open
atmosphere; the amount of energy needed for this to happen is called the metal’s work function. In the microwave, this energy is
supplied through the addition of heat, which is known as thermionic emission. Many metals exhibit this effect, but none have the
other properties of tungsten that make it essential for the microwave oven’s magnetron. Tungsten has the distinction of having the
highest melting point of common metals. This is useful in that you can run the magnetron at higher voltages and produce a larger
number of electrons than a metal that melts at a lower temperature. Yet, because of tungsten’s high work function, pure tungsten
would operate at too high a voltage, requiring too much power. To reduce the required power, yet keep the high melting point of
tungsten, engineers make microwave oven magnetrons from a combination of tungsten and thorium. Thorium has a much lower work
function than tungsten and so will emit electrons at a lower voltage, but on its own would melt at the temperatures that result from the
necessary voltages to operate the unit. In a mixture of the two, the thorium can be turned into thorium oxide at the electrode’s
surface, then as the surface thorium evaporates during operation, it is continuously replenished by diffusion from inside the
electrode.
If the magnetic field is such that the circulating electrons just touch the cavity opening, an oscillating electromagnetic wave
will be formed. Left The lines indicate the electric field created by the cavities. They are shown at one instant in time; these
fields reverse direction periodically as the magnetron operates. These oscillating fields influence the path of the electron. Right
They create a “spinning pinwheel of electrons” when combined with the magnetic field perpendicular to the plane of the paper.
(The ends of magnetic field “arrows” are shown as circles in the cavities.) Shown within a pinwheel spoke is the path of a single
electron as it spirals toward the anode. This pinwheel rotates, creating an electromagnetic wave with a great deal of power.
Fourth, once the oscillations from the resonators have been established, the electrons maintain the
oscillations that would otherwise die out as the result of joule heat losses in the copper block. A critical
step, and the step that gives the magnetron high power, is that the RF fields from the cavities draw energy
from the orbiting electrons. In other words, oscillations are maintained at the expense of the energy of the
orbiting electrons. In a magnetron with no cavities, the electrons would cycle back toward the cathode (if
the magnetic field were strong enough). When this rotating wheel induces oscillations in the cavities,
oscillating magnetic and electric fields extend from the cavities. These fields speed up or slow down the
electrons. When they speed it up, these electrons just fly back to the cathode because the RF field increases
their curvature. In fact, those electrons that absorb energy—those that speed up—from the R F field are
eliminated at once. But a number of electrons are slowed down, and their curvature is decreased.
These slowed‐down electrons will eventually reach the anode and thus contribute to the current and
give some of their energy to the RF field. This results in a distortion of the rotating cylindrical sheath that
occurred when no cavities were present; the sheath now appears as a smaller cylinder with four spoke‐
like ridges running parallel to the axis. This pinwheel rotates with an angular velocity that keeps it in step
with the alternating RF charges on the anode segments, and the ends of these spokes may be thought of as
brushing the ends of the segments and thus transferring charge from cathode to anode.
Why Do the Cavities Resonate?
It seems counter‐intuitive that a simple hole like the cavities would cause electrons to resonate, but the
phenomenon can be understood by modeling the system as a grandfather clock.
A pendulum moving back and forth, as in a grandfather clock, is in resonance. This means that it moves
back and forth at a specific frequency. Resonance occurs when a system is able to store and easily
transfer energy between two or more different storage modes: In the case of the pendulum, this is the
change from kinetic energy to potential energy. Resonance isn’t limited to mechanical systems; it can also
occur in electrical systems by storing energy in magnetic and electric fields. The latter are analogous to
the two energy storage modes—kinetic and potential energy—of a grandfather clock.
As the pendulum’s bob reaches the outer edge of its arc, its speed drops to zero. At this point, its energy is completely
potential, but it gives up that potential energy in exchange for kinetic energy as it begins its descent back to the other side of its
arc. At the very bottom of the arc, the pendulum is moving at its fastest speed; all its energy has been converted to kinetic
energy. It gradually turns this kinetic energy into potential energy as it reaches the limit of its arc on the other side. A key point
to note from this example is that systems have specific resonance frequencies; once the pendulum begins to swing, it will swing
at only one frequency (typically 60 HZ) until it dies out because of wind resistance.
In an electrical circuit, the two modes of energy storage that create the oscillation are electrical and
magnetic fields. A simple oscillating circuit contains an inductor and a capacitor.
When current flows through an inductor, the inductor stores the energy as a magnetic field, and the
capacitor stores energy in a electric field.
Left An inductor stores energy in a magnetic field. The energy from a current flowing through the coil is stored in the
magnetic field generated by the coils. If it were short‐circuited, it would release its stored energy as current. Middle A capacitor
consists of two separated plates. As current flows, charge builds up on the plates until the capacitor is full. If the capacitor were
short‐circuited, it would release its energy, creating a current. Right If we combine the two elements, we make what’s called an
LC circuit—L is the symbol used for inductance, and C for capacitance. As described in the next figure, this circuit can oscillate.
Store is a key word here: It implies a time‐varying element. Let current flow through an inductor, and it
will store the current’s energy in a magnetic field until full. Then, when the current is shut off, the
inductor’s magnetic field slowly dissipates as current. In a capacitor, something similar happens: When
current flows through the capacitor, it stores a charge that is released when the current stops. At its
simplest, an inductor is just a tightly wound coil of wire. As we’ve learned, current flow creates a
magnetic field. The coil of the inductor concentrates this magnetic field. A capacitor consists of two metal
plates held apart a short distance. When current flows, charge builds up on each plate and creates an
electric field. If we place an inductor and a capacitor in a circuit and charge only the capacitor, this
circuit will resonate.
You might ask whether such a resonant circuit can be used to generate radio frequency waves, since all
we need to generate electromagnetic waves are oscillating magnetic and electric fields. Indeed, this does
work; in fact, the earliest radios used circuits comprised of capacitors and inductors modified by a
vacuum tube or transistor to keep the circuit oscillating. The latter two elements did the electronic
equivalent of mechanically tapping the pendulum of a clock when its oscillations began to decay. So, it
would seem that this would be a great way to make R F in the microwave range, and it does work to a
degree. We can produce radio frequencies but eventually, at very high frequencies, the rapid back‐and‐
forth motion burns out the circuit elements.
For high‐frequency radio waves in the microwave regions, we can build the circuit into the copper
anode without any wires or apparent circuit elements. A cavity alone does the trick. In the anode,
engineers cut cylindrical cavities. Viewed from the top, it looks like a loop of wire where the ends are
very narrow. If current flows around the loop, the cavity causes the cylinder to behave like an LC circuit.
The sides of the narrow slot are so close together that they act like a capacitor—it stores energy as charge
builds up. This behavior is completely analogous to the resonator created by blowing over the top of a
bottle.
As the capacitor discharges, its current flows into the inductor. Because of the flowing current, the inductor generates a
magnetic field that stores the energy contained in the current. Eventually, the capacitor becomes completely discharged.
Current no longer flows, so that inductor begins to release its energy as current flows, which causes the capacitor to become
charged. Once the inductor is completely discharged, current no longer flows, and the capacitor then releases its energy,
starting the cycle over again. Shown underneath every step is the energy stored in the inductor or capacitor. The total energy
moves back and forth between the inductor and capacitor to store the energy, just as a pendulum’s energy oscillates from
potential to kinetic. The rate at which this exchange happens in the circuit (the resonant frequency) would depend on the size of
the inductor and capacitor. Of course, the oscillations cannot go on forever. Just as in a pendulum clock, the resonance decays.
The energy being shuttled back and forth is lost as heat in the wire, the inductor, and the capacitor.
In this cavity, current moves around the circular opening. The gap creates a capacitance. The energy transported by the
current gets trapped in the magnetic field created, making the cavity act like an inductor and setting up an oscillation just like in
an LC circuit. The solid arrows indicate the electric field, and the dotted arrows show the magnetic field.
Primer: Electrons, Energy Levels, and Light Emission
AN ATOM OR MOLECULE consists of a charged nucleus or nuclei surrounded by a cloud of electrons --
a cloud because really we can only talk about the probability of an electron’s location.
Bonds form between atoms when electrons are shared. While we have a commonplace image of an electron as a small
sphere orbiting a nucleus, it’s best to think of the electron as a cloud of probability. In the molecule on the left, you can see that
the electron is more likely to be found near the nuclei than far away. A crystal forms (as shown on the right) when many atoms
share electrons to form a lattice. Near the nucleus, the cloud is the most dense, and as we move away from the nucleus, the
electron density tapers off.
Ground states and excited states. Imagine a single electron from that cloud. That electron will be
attracted by the positive charge of the nucleus and so it will desire to move toward it, but it will be
repelled by the other electrons clustered around the nucleus. Thus, it will find some place to “rest” that
balances these two forces.
Shown on the left is an electron in its ground state: It sits on the probability cloud at its most likely location. We can add
energy to move this electron further away from the nuclei. This is called an excited state or an excited electron. Typically
chemists represent these two states on a diagram as shown on the right. These electrons will eventually return to their ground
states. We could, of course, move an electron so far from the nuclei that it leaves the molecule forever. That is called ionization.
We could do work and add energy to the electron, and that would move it away slightly from the
nucleus. We would call this an excited state of the molecule or say that the electron is excited. Naturally,
the electron will eventually return to its lower energy position, called the molecule’s ground state, once
we stop supplying energy to it. Think of a playground swing: Its natural state -- ground state -- is just
hanging underneath the bar. If we add energy by pushing it, we can create an excited state. With that
energy we can do many things, like knock something over in its path. However, if we use the energy to do
that, the swing loses its energy quickly. Regardless of what we do, the swing will eventually return to its
ground state: All the energy will be lost as heat, either in the air resistance of the swing, or in friction
between the swing’s chain and the metal support bar. Like a swing in motion, there are many ways an
excited electron can return to its ground state, but the electron has an odd property not apparent in the
swing’s motion.
Quantization. You would think that an electron could occupy any value of energy around the atom.
You might guess that we could add whatever energy we want and move the electron any distance from the
nucleus, but we cannot.
Let’s return to our familiar example of a grandfather clock. When a pendulum moves back and forth, it
sweeps continuously through every section on its arc; it moves smoothly without jerks, changing its energy
slightly on each part of its arc. Yet that motion is only an illusion; in reality the pendulum moves in
incredibly small, piecewise steps, jumping from one position to the next, never existing in between. We
call this motion quantized: a series of discrete places -- or, more precisely, energy levels -- where the
pendulum can be. To our eye, the pendulum's motion appears continuous because the difference between
the energy levels is far too small for the human eye to detect.
According to quantum mechanics, the pendulum can only have energy values -- and thus locations on
its arc -- of: (1/2)hν,( 3/2)hν, (5/2)hν and so on. Here h is Planck’s constant and ν is the frequency of the
pendulum, typically 2 cycles per second for a grandfather clock. Planck’s constant is tiny: 6.626 x 10-34
joule-seconds, where one joule of energy is about what you’d feel if you dropped this book on your head
from about a foot above your head. The tiny value of Planck’s constant means the spacing between energy
levels associated with the possible locations of the pendulum are very small. In our everyday world, the
energy levels associated with the pendulum are indistinguishable from one another, but when we get down
to the size of atoms, these effects become important.
At the microscopic level, then, only discrete and quantized energy levels are available, hence the terms
“quantum” and “quantum physics.” In the drawing of energy levels in the figure above, there is nothing in
between the two levels. This isn’t just a peculiar phenomenon of electrons in atoms; at its core, our world
is completely quantized, even in terms of motion.
Why is the world quantized? For the purpose of this book, we will accept the following fact: At a
small enough scale there are discrete energy levels to entities such as electrons. We won’t bother with
anything beyond this, we’ll only ask the technological importance of it.
Decay of excited electrons. We can excite electrons by adding energy -- heating an object until it
glows as in heating a filament, passing electricity through it like a fluorescent lamp, injecting current as in
an LED (light emitting diode), or using the energy from a bright light, as in a ruby laser. All these ways to
add energy are analogous to pushing a swing to “excite” it. An excited electron decays, losing energy by
returning to its ground state, in several ways. It can do so in a way that emits no light, called non-radiative
decay, where instead of light, it gives off heat. To be specific, it can release energy by vibrating the
molecule or by inducing a collision or rotation. The electrons returning to lower energy could also cause
the molecule to dissociate or break up. Lastly, the excited electron could cause the molecule to
photochemically react. But most important technologically is a radiative loss of energy; that is, the
returning energy being released as a burst of electromagnetic radiation.
There are many ways to induce this burst. Of main importance for a laser is using light or electricity to
create the excited states. If we use current, as in an LED, it’s call electroluminescent. There are two ways
that an electron can give off light: spontaneous emission, which results in incoherent light, or stimulated
emission, where many electrons radiate together, creating light that is in phase. The former occurs in light
bulb filaments and glow-in-the-dark toys, the latter happens in a laser.
Chromium, Helium, & Neon
[kroh-mi-uhm] | Symbol: Cr | Atomic number 24 | Atomic weight 51.9961 amu
[he-lee-uhm] | Symbol: He | Atomic number 2 | Atomic weight 4.002602 amu
[nee-on] | Symbol: Ne | Atomic number 10 | Atomic weight 20.1797 amu
In this chapter, we highlight the role of elements used in the creation of laser light. The
first laser, which used chromium, appeared in 1960, followed rapidly by those based on
helium and neon.
The word chromium comes from the Greek chromos, meaning color. Indeed, the oxide
of this silvery white-gray metal, discovered in a Siberian gold mine in 1766, creates vivid
colors in a variety of compounds. For example, the mineral alexandrite contains a small
amount of chromium which absorbs light over a narrow range of wavelengths in the
yellow region of the spectrum. Alexandrite from Russia’s Ural Mountains (it is named
after the Russian Tsar Alexander) looks green in daylight but red by incandescent light.
Ruby is largely a crystal of aluminum and oxygen, where chromium replaces a small
fraction of the aluminum atoms. The chromium strongly absorbs yellow-green light,
which makes the gemstone glow red when a blue light shines on it. This red glow forms
the basis of a ruby laser.
The inert duo of helium and neon revolutionized lasers. Both are colorless, tasteless,
and odorless gases. Neither reacts with anything strongly; in fact, neon won’t even form a
compound with itself, a rare feat for gases. Helium’s discoverer detected it in the sun long
before anyone found it on the earth. Using a prism, French astronomer Jules Janssen
observed a yellow band in the light coming from the sun. He surmised, correctly, that the
line corresponded to a new element. British astronomer Norman Lockyer discovered
helium at the same time, and he named it in honor of the sun, basing it on helios, the
Greek word for sun. Stars, such as the sun, produce huge amounts of helium by the
fusion of hydrogen. No surprise, then, that it’s the second most common element in the
universe after hydrogen (the two together account for 99% of all observed matter).
In contrast to helium’s extraterrestrial discovery, neon became widely available for the
most human and earthbound of reasons: war. At the start of the nineteenth century, a
Frenchman, Georges Claude, searched for a way to recover oxygen from air. What really
ignited demand for oxygen was an arms race: The steel industries of Europe needed lots
of pure oxygen to create extremely strong steel for armaments. By freezing air, Claude
was able to extract pure oxygen, but since air contains more than just oxygen, he also
recovered the rare gas neon. Neon gas was so expensive at the time that, prior to Claude’s
work, it was used only for exotic purposes. For example, in 1897, a special light display
was made using neon and other rare gases to celebrate Queen Victoria’s Diamond Jubilee.
Today, we see it all the time: The glow inside these neon lights is a sort of tamed lightning
used in displays everywhere.
Electricity strips electrons from neon, and the electrons then smash back into the
positively-charged remains, producing a brilliant light. In combination with helium, neon
produces one of the world’s most useful and inexpensive lasers.
How a Laser Works
ON JULY 21, 1969, astronauts Neil Armstrong and Buzz Aldrin walked on the moon. Although getting to
the moon was a “giant leap for mankind,” few know that one of their tasks for this pioneering mission was
to install a small metal plate covered with 100 reflectors made of fused quartz on the surface of the moon.
Once the reflectors were placed there, geo-scientists made highly accurate measurements of the
separation of the moon and the earth. From these measurements, they learned fundamental facts about how
the moon behaves; they learned that it spirals away from earth at a rate of 38 mm a year, that the moon has
a liquid core that is 20% of the moon’s radius, and they even showed that Einstein’s Theory of Relativity
predicts the moon’s orbit with incredible accuracy. They measured the moon-earth distance by firing a
high-powered laser towards the reflector and measuring the time it took the beam to travel there and back.
Using the speed of light, they calculated the distance between the earth and the moon to an accuracy of 2
cm, an astonishing accuracy considering the moon is 385,000 km from earth. The amazing part of this
story isn’t only humans walking on the moon for the first time or the accuracy of the measurement, but that
we could, from earth, shine light onto the moon!
This exact measurement was made possible by the invention of laser light. Laser -- an acronym for
Light Amplification by Stimulated Emission of Radiation -- produces light from a phenomenon known as
stimulated emission, which we’ll explain later in this chapter. Although lasers are commonplace today,
they are different from normal light in a few ways. Imagine using a flashlight to illuminate the moon: Its
beam would dissipate into the atmosphere long before it reached its destination. Though powerful, the
flashlight lacks the three key characteristics of a laser that allow light to make it all the way to the moon
and back and still be measurable.
Narrow beam. A key characteristic of laser light is an incredibly focused, narrow beam of light. When
the light travels to the moon, the beam spreads out to a diameter of 7 km, but without beginning as an
incredibly narrow beam, it would diffuse completely in the earth’s atmosphere.
Great intensity. Since all of the light’s energy is packed into a narrow beam, it is bright enough to
reach the moon, be reflected, and return to earth.
Monochromatic beam. Its single pure color allows it to be detected when it returns to earth: Scientists
can separate the return signal by narrowing their search for the return beam to a nearly single color of
light.
Using a laser to reach the moon is an exotic, even esoteric, use of a laser, but it highlights how the
union of these three characteristics gives lasers immense technological importance. Lasers flash messages
through fiber optic lines, read grocery bar codes, and even heal people. Eye surgery, for example,
highlights again the importance of all three of a laser’s unique properties. Surgeons use green lasers
because that color passes through the eye’s vitreous humor without being absorbed and thus does not
cause damage. The laser beam then strikes the retina where its great power can “weld” a detached retina
back in place. The beam’s narrowness lets the surgeon affect only the area of the retina that needs to be
repaired. Without the three unifying characteristics of lasers, doctors wouldn’t be able to perform such
surgery.
Key Characteristics of a Laser Beam
By comparing a laser’s light to that of a light bulb, we can refine the characteristics that make a laser
unique.
Engineers typically recast the three key characteristics of a laser mentioned above in more technical
terms. Specifically, unlike incandescent light, a beam of laser light has:
Temporal coherence . All of the peaks and troughs in the waves of light line up. This gives the light
beam an astonishingly pure color, best characterized as almost a single wavelength of light. An
incandescent light shines with a variety of wavelengths.
Spatial coherence. All of the light rays are evenly and equally spaced. This gives the beam great
power: In an incandescent light, the intensity varies -- in some places its rays are packed tightly,
elsewhere very spaced out.
Collimation. All of the rays from the laser are parallel. This means the beam contains only a small
area, giving the user tight control in, say, removing a tattoo or reading a bar code. In an incandescent light,
the light rays scatter in all directions, resulting in a beam of overall lower intensity.
A light bulb produces incoherent light: it is made of many wavelengths, travels in all directions from the bulb’s filament, and
varies in intensity. Laser light is characterized by coherence -- meaning that the light exiting the laser is in phase -- of a
very narrow band of wavelengths and travels in the same direction.
How does a laser create this amazing beam of light? For that matter, how does anything create light?
Why Does Anything Give off Light?
The basis of all human-made generation of light lies in the movement of electrons within an atom or
molecule. If we supply energy -- an electrical current, heating, or even a flash of powerful light -- we can
promote electrons in a molecule from their normal low-energy state to higher energy. Typically, an
electron persists in this excited state for only a short time before it returns to lower energy. When it
returns, it must give off that extra energy. Most often it just dissipates energy as heat, but sometimes it
releases that energy as a burst of light.
In a typical incandescent light bulb, about 90% of the energy supplied dissipates as heat, rather than as
visible light. A current passing through a tungsten filament heats the tungsten, promoting some of the
filament’s electrons to higher energy. As the electrons return to lower energy, some give off light,
although most dissipate this energy as heat (think of how hot a light bulb gets after it has been on for
awhile). However, heat isn’t the only way to “excite” electrons. For instance, if you shine a blue laser
pointer on a piece of ruby, it’ll glow a robust red. It’s absorbing the blue light, using its energy to promote
electrons to higher energy, which then decay, releasing a brilliant red color. A similar phenomenon
occurs with a glow-in-the-dark football. After being held in the sunlight for half an hour, it’ll glow
brightly in a dark room. The zinc sulfate copper compound embedded in the football absorbs the higher
energy sunlight and then slowly releases it as an eerie green. The sparkle from an irradiated ruby and the
glow of the football lie at the core of how a laser works.
Lasers & Spontaneous Emission
In some ways, a laser is like a glow-in-the-dark football on steroids. The key difference in how
sparkling rubies and glowing footballs make light and how a laser makes light lies in the number of
excited electrons. The difference isn’t how it’s done, but rather how many times it happens. In a laser, an
intense burst of light creates many excited electrons; their decay gives the laser beam its unique
properties.
To make a laser, we need a medium with at least three energy levels: a ground energy level and two
higher energy levels. With a burst of energy, we can create a population inversion; that is, more electrons
in the highest energy level than in the energy level just below it. This inversion initiates a most amazing
thing: When one of the electrons returns to a lower energy state (releasing its energy as light), that light
ray triggers other excited electrons to release their energy as light -- and here’s the fascinating part -- in
phase with the first ray of light released. This stimulated emission gives the laser beam its great
coherence. Contrast this with the emission of light by the football: There, the excited electrons decay by
spontaneous emission, giving off light rays that are not in phase; to use technical language they have no
temporal coherence. By creating the population inversion, we dramatically increase the chance that
stimulated, rather than spontaneous, emission will occur.
To consistently and reliably create the stimulated emission at the core of a laser, to put it into practice,
requires a great deal of engineering.
In a glow-in-the-dark toy, the electrons start a) in the lowest energy level, called the ground state, then b) light promotes or
excites a few of the electrons to a higher level. These excited electrons drop c) to an energy level between the excited and
ground states. The electrons from that level can return to the ground state, giving off light. This spontaneous emission creates
incoherent light. In a laser, the electrons also start d) in the ground state, then e) a huge burst of energy rapidly promotes many
electrons to the excited state. The excited state electrons drop to an energy level between the ground state and the excited
state. When these electrons f) return to the ground state, they give off coherent light because of the population inversion; that
is, there are more electrons excited than in the ground state.
Why Does Stimulated Emission Occur?
The laser requires quantum mechanics for a complete description because spontaneous and stimulated emission have no true
classical analogies. To deeply understand laser technology requires the use of QED (quantum electrodynamics), which is a step
beyond the quantum taught to most engineers. Stimulated emission seems like a very odd concept, yet it is the spontaneous emission
of light (light that isn’t in phase, like the kind that comes from a glow-in-the-dark football) that truly reveals the strangeness of the
microscopic world. Understanding spontaneous and stimulated emission brings the wave-particle duality of light to the forefront of
our thinking and highlights its essential “weirdness” in relation to our macroscopic experience of the world. At a very simple level,
one can visualize stimulated emission as either a particle or a wave phenomenon.
A photon, behaving like a particle, ricochets back and forth throughout the laser cavity. Like a billiard ball, it hits electrons in
their excited state, inducing them to return to a lower energy level and release some of their energy as photons. In that collision, the
impacting photon loses no energy -- it’s an elastic collision similar to what happens to idealized billiard balls. The photon released in
turn collides with another excited electron to release more photons that eventually create the great cascade that gives the laser its
powerful light. An alternative description to this cascade can be given by viewing stimulated emission as arising from the wave
nature of light.
Picture the light racing back and forth within in the laser as a wave. The excited electrons act like antennas detecting the
fluctuations of the electric field, which push and pull until the electrons release a wave of light with a phase that is in complete
sympathy with the first wave of light.
Although greatly simplified, these two descriptions make sense in relation to our own experience of the world and help us
understand how lasers work. Spontaneous emission, though, is much harder to square with our experience of the world.
Spontaneous emission occurs even if there are no photons or light waves (choose your poison) to entice a light from an excited
electron. To understand why spontaneous emission happens, we need to appreciate the fundamental way in which the quantum world
differs from our human-sized world. Spontaneous emission occurs because on a small enough level, the world is never completely
still. Werner Heisenberg showed in 1927 that the position and energy of an object can never be exactly defined at the exact same
time. This means that nothing ever stops moving, a notion foreign to our commonsense experience. For example, we think of a
vacuum as the absence of everything, yet microscopically it is a hot-bed of activity. The vacuum contains a small amount of energy,
called zero-point energy, that can induce spontaneous emission. If you find that hard to wrap your head around, keep in mind that
there is no analogy from the human-sized world to guide you, so it really makes no classical sense.
The first ruby laser was only a few centimeters long. Its center was a ruby cylinder mirrored at both ends; one of them,
though, was partly silvered so eventually light could escape. A high-powered flash lamp surrounding the cylinder provided the
energy to create a population inversion.
He coated the ends of the cylinder with silver to make them mirror-like, leaving a small hole in the
center of one mirrored end to allow a small amount of light to escape. Maiman flashed the lamp for a few
milliseconds; the intense burst of light caused electrons in the ruby to become excited, although most of
the flash was lost as heat. The ruby rod emitted bursts of incoherent red light via spontaneous emission.
He kept flashing the light, effectively “pumping” the electrons, and eventually he created a population
inversion. The radiative decay of an excited electron then triggered stimulated emission, releasing an
avalanche of light with all of its waves in phase. The light swept back and forth across the ruby cylinder,
reflected by the mirrored ends. These repeated reflections both narrowed the range of wavelengths of
light emitted from the ruby and eliminated any rays not parallel with the ruby cylinder’s axis. This creates
the highly coherent, monochromatic beam characteristic of a laser.
Maiman succeeded in creating the world’s first laser and so was twice nominated for the Nobel Prize.
He never won; instead the Swedish Committee awarded it to the inventors of the MASER, which paved the
way for lasers.
Maiman’s laser had several imperfections that made it non-ideal for portable, everyday use. For
instance, it required a lot of energy and gave off large amounts of heat. Ruby lasers could only be easily
used in pulsed-mode, rather than continuous, because of the intense heat built up. To make the laser a less-
specialized, everyday tool required other lasing media than ruby. Of greatest importance are gas and
semiconductor lasers.
Helium-Neon Lasers. By replacing ruby with a gas mixture, engineers create lasers than can be
operated continuously simply by removing heat using passive air cooling at low powers or, for high-
powered operation, by using forced air from fans. A stellar example of such a laser is the helium-neon
laser, which produces very little waste heat, is easy and cheap to construct, and is very reliable. A simple
flick of a switch, and a red beam of wavelength 632.8 nm emerges.
At the core of the helium-neon laser lies a sealed glass tube filled with helium and neon with reflecting
ends, as in the ruby cylinder. Its inventors replaced the cumbersome flash lamp of the ruby laser with a
large jolt of electricity: zapping the gases with 9,000 volts, creating excited electrons in the helium gas.
These helium molecules in turn collide with neon gas molecules. The energy levels of the exited helium
overlap with the exited states of neon, so the collisions transfer energy to the neon, creating excited
electrons in the neon. The stimulated emission of these electrons radiatively decaying gives off lights of
wavelength 632.8 nm (a red light) and infrared light of wavelengths 1150 nm and 3390 nm that human
eyes cannot detect.
A helium-neon laser contains a mixture of both gases. A voltage discharge excites the electrons in the helium atoms that,
through collisions, transfer their energy to the neon atoms. (Helium and neon were chosen because their energy levels in the
excited state match up closely.) The electrons in the neon decay to give off light. There are three main transitions, two in the
infrared and one in the visible (red) part of the spectrum.
Why Use Two Gases in a Helium-Neon Laser?
Neither gas alone would make a very good laser. The excited helium atoms are very stable in the sense that the electrons stay
excited for a long time but cannot easily radiatively decay. In contrast, the neon won’t stay excited for long, but it does radiate
easily. The combination of the two uses the best properties from both gases.
A gas laser has the huge advantage over the first ruby laser in that it doesn’t produce as much heat. The
radiative decay of the excited neon atoms is from one higher-energy state to another, rather from an
excited state to the ground state as in a ruby. A helium-neon laser can therefore operate at a lower
temperature than a ruby laser can. Recall that at the quantum level, the temperature of a substance reflects
how the electrons are distributed among a molecule’s energy levels: A higher temperature means more
electrons in higher energy states. So, in a three-level system -- that is, three energy levels -- like ruby, we
need to promote a large number of electrons from the ground state to a higher energy state to get a
population inversion. (The key population inversion is between the ground state and some higher energy
level.) In a four-state system like helium-neon, we only need to get a population inversion between two
higher energy states. The radiative decay occurs when an electron decays from one higher energy state to
another high energy state, rather than to a ground state. This means that we can have more electrons in the
ground state for helium-neon than for ruby and thus operate at a lower temperature. You might also think
of it this way: We need to promote more electrons in a three-level system than in a four-level system.
Each of the excited electrons generates heat -- they make the crystal vibrate -- as it drops to a lower level.
More electrons thus means more heat.
To achieve a population inversion in a three-level system, more electrons need to exist in energy level 2 than in the ground
state, energy level 1. In a four-level system, the population inversion necessary for stimulated emission needs to exist only
between levels 2 and 2’. This means that we don’t need to promote as many electrons in a four-level system (like HeNe) than
we do in a three-level system (like ruby). This is what allows a HeNe laser to operate cooler than a ruby laser.
Who Really Invented the Laser?
The headline on a 1965 New York Times story announced -- on page 96 -- a possible problem for the rapidly growing laser
industry: “A Laser Patent Fought in Court: Former Graduate Student Seeks Rights to Device.” At stake was millions in licensing
revenue. The plaintiff, Gordon Gould, claimed to have invented the laser while working as a graduate student at Columbia
University; he had failed to get his doctorate there because of his preoccupation with laser research. He even noted that he and his
wife separated because of his obsession with lasers. Gould’s lawyer claimed that Gould conceived the laser in his Bronx apartment
in 1957, and then had a notary public in a nearby candy store notarize his notes and drawings. Gould didn’t file for a patent,
though, because he thought you had to have a working model.
At nearly the same time Charles H. Townes, a professor at Columbia, and his post-doctoral researcher Arthur L. Schawlow
developed the same idea. They filed for a patent in July of 1958. At issue was the parallel mirrors that cap the ends of a resonant
cavity -- a key, perhaps the key, to operating a laser. Eventually Townes and Schawlow secured a patent -- and separate Nobel
Prizes -- for their work. Gould received a relatively unimportant patent for laser production of x-rays in 1968. The 1965 lawsuit,
though, started Gould on a three decade quest to gain priority for inventing the laser.
He fought and fought until by 1977 he finally received a patent for the optically pumped laser amplifier; that is, the heart of
every commercial laser! Soon thereafter Gould was also awarded nearly 50 patents, which covered most of the laser design used
commercially. Armed with this portfolio of patents Gould teamed with financial backers to campaign for all manufacturers to pay
royalties on nearly all lasers. Even though Gould gave away 80% of his profits to his backers, he still earned, according to Nick
Taylor, an expert on the history of the laser, “millions upon millions of dollars.”
Resonator Cavity
While the light generated by the ruby or the helium-neon gas (otherwise known as the gain medium)
seems the most exciting part of a laser, the real key to making the precise wavelength as well as the great
collimation of laser light lies in the design of its cavity. When a laser first begins operation, the
wavelength of light generated in the gain medium is not very sharp by laser standards. For example, when
a decaying electron in neon emits light in a HeNe laser, that light consists of wavelengths ranging from
roughly 620 nm to 640 nm. While this difference of 20 nm would be indistinguishable to our eye, it would
destroy the usefulness of a laser. For a typical HeNe laser, the range of wavelength is an amazing 0.01 nm
-- from 632.79 to 632.8 nm! Without the cavity, not only would the range of colors be broad by laser
standards, but the light would not be highly collimated; that is, its rays would go in all directions. The
resonator cavity, the area between the mirrored ends of the ruby cylinder or tube containing helium and
neon, both narrows the range of wavelengths and collimates the beam by causing the light’s rays not
parallel to the cavity’s axis to destructively interfere and die out.
The repeated reflections inside the cavity create a standing wave. (See the primer on waves that
proceeds a discussion of the the microwave oven for background on standing waves.) This means that
only light of a particular wavelength can exist in the cavity. Specifically, only light with a wavelength
such that its “nodes” -- where a waveform passes the x-axis -- have a value of zero at the mirrored ends
of the cavity, or light whose half-wavelength is an integral multiple of the cavity’s length. An infinite
number of these wavelengths can exist in a cavity, all spaced equally in energy. So, if we choose the
cavity size correctly, only one or a few of these allowed wavelengths will exist. When these allowed
wavelengths are convoluted with the slightly broader range of wavelengths emitted by the radiative decay
of the gain medium, the result is a narrow single wavelength.
Shown on the left is the resonant cavity of a laser. An infinite number of wavelengths can exist inside the cavity as long as
they are an integral number of half-wavelengths. Shown above are three of those wavelengths: λ, λ/2, and 3/2λ. (Note that these
lines are spaced equally in energy). This spacing depends only on the cavity size and not the wavelength of the laser’s light. The
right shows the light output from the helium-neon red line: Its width indicates a range of color or, better put, wavelengths. The
cavity only allows the waves shown to exit. For this 5 cm HeNe tube -- much shorter than a real tube -- three narrow laser lines
would appear: Those at λ, λ/2, and 3/2λ, with the λ wavelength the most intense. (Here the wavelengths are drawn schematically
for clarity.)
In engineering such a cavity, the most critical part is the degree to which the mirrors are parallel. If
they deviate by more than a few percent, the cavity will allow all sorts of wavelengths to exist within the
cavity and destroy the monochromaticity of the laser. In practice, this means the mirrors cannot deviate
from being parallel by more than one-third of the operating wavelength. For the first ruby laser, Maiman
built his mirrors so that at no point did the distance between them differ by more than 694.3 nm/3 or
231.43 nm from top to bottom.
Semiconductor Lasers
Although ruby lasers proved that a functional laser is possible, and HeNe lasers illustrated their
practical usefulness, of greatest importance in our world today is the semiconductor laser. They’re used
in laser pointers, bar code readers at grocery stores, and in laser printers. Their key advantage over
solid-state or gas lasers lies in the fact that they are compact and cheap because they can be made with the
automated methods used to manufacture computer chips.
Most semiconductor lasers contain the same basic components as the ruby or helium-neon laser,
although the method of producing light comes from electroluminescence. That is, a current drives the
stimulated emission, not a flash lamp or voltage discharge. A junction of two different types of
semiconductors replaces the ruby or helium-neon mixture. Two edges of this “sandwich” are surrounded
by material of a higher index of refraction, so no light can exit the sides. The other two ends are highly
polished to create the familiar resonant cavity. When current is injected into the junction, light appears as
a colored band of light emitted from the area.
A typical simple semiconductor laser. Light is made at the junction between an n-type and a p-type semiconductor. The sides
of the laser are of high refractive index so no light escapes; the ends are highly mirrored to create a resonate cavity.
The two materials making up the junction are a p-type and n-type semiconductor. p-types transfer
current by positive carriers called holes, and n-types by negative carriers called electrons. (See the end
of the chapter for a section on semiconductors.) The junction of these two semiconductor types creates a
diode; that is, the device allows current from a battery to flow in only one direction. The transfer of
current occurs at the junction between the two types of semiconductors: When the battery injects electrons
in the n-type semiconductor, the diode allows electrons to flow out at the junction: The electrons
recombine or neutralize with positive holes. When they recombine, they can give off light.
The energy difference between a separated electron and hole and their combination is analogous to the
discrete energy levels in a ruby or helium-neon laser. The energy of the separated pair is higher than when
they are together; therefore, when they combine, they release energy. Just as in a ruby or a mixture of
helium-neon, the energy can be released as heat -- the vibration of the semiconductor crystal -- or it can
give off energy as a burst of light. Rather than create a population inversion by a flash lamp, energy is
added by passing current through the diode. By “injecting” a large amount of current flowing through the
diode, we create many electrons that can then combine with holes, giving off light and inducing stimulated
emission. The edges of the semiconductor sandwich are highly polished to collimate the laser.
Erbium Amplifier for Fiber Optics Cables
Understanding stimulated emission allows us to see how one of the cleverest engineered objects of all time works. It’s called an
erbium amplifier, which is used in fiber optic cables. To understand the revolutionary nature of these devices you must understand
something about sending signals through fiber optic cables.
A signal is encoded as flashes of light that are then sent into a fiber optic cable. Via total internal reflection they “bounce” down
the cable. While cables do a great job of transmitting the light pulses, the light eventually escapes out the walls of the fiber.
Engineers call this the attenuation of the signal; typically a signal will travel many miles until it has dropped to about 1% of its
original strength. At this point it needs to be amplified or it will be lost forever. For example, TAT-8, the first undersea cable across
the Atlantic Ocean that connected Tuckerton, New Jersey with Widemouth Bay, England, and Penmarch, France, used amplifiers
every 30 miles. These repeaters were complex and energy intensive units that used photocells to convert the light to an electrical
signal, amplified or boosted that newly transformed signal, and then converted it back to light so it could continue 30 miles to the
next repeater. These repeaters were expensive, cumbersome, and, being complex, a source of potential failure. An erbium amplifier,
called an EDFA, boosts the light using stimulated emission, eliminating the need for converting the optical signal to an electronic
one.
Here’s how the amplifier works. Erbium has a rich electronic structure; that is, it has many energy levels that will emit light once
its electrons are excited. Inside the glass fiber optic cable engineers build sections, every 30 miles or so, that are “doped” with
erbium. Using an infrared laser, they can populate the higher energy levels of erbium and create a population inversion. When the
optical signal arrives, its photons induce these excited electrons to decay – recall that this is stimulated emission -- and this burst of
light adds to the incoming signal thus boosting it optically. The emission is stimulated so the photons produced are exactly in the
same phase and direction as the amplified signal, and, importantly, the signal is amplified along its direction of travel only. The
greatest advantage of these erbium-doped amplifiers over electronic amplifiers lies in multiplexed signals.
To increase the capacity of a fiber optic cable, engineers send multiple messages simultaneously. They do this by encoding each
signal as a different wavelength of light. (Essentially each signal is a different color, but the wavelengths are so close together that
our eyes could not tell the difference.) Picture these signals traveling down the the fiber optic cable, slowly losing intensity until they
need to be amplified. As they enter the erbium enriched section they are automatically boosted individually. The incoming light of a
particular wavelength matches exactly one of the energy splittings in erbium, another signal matches another wavelength.
Lasers began as the stuff of science fiction. First described in Alexey Tolstoy’s prescient 1927 novel
The Hyperboloid of Engineer Garin as a “filament of light of any thickness desired,” the laser was used
as a death ray. The laser’s exotic and supposedly deadly properties stayed a staple of science fiction TV
shows and movies even after its invention. Yet with age, it mellowed to assume the most commonplace of
tasks -- like quietly, unintrusively scanning groceries.
In Depth: Semiconductors, Electrons & Holes
Recall that there are metals that conduct electricity freely, insulators (like rubber) that don’t allow electricity to flow, and semiconductors
that conduct electricity but not as well as metals. What’s most important about semiconductors is that, unlike metals, semiconductors can
conduct electricity by two mechanisms: the movement of negative or positive charge carriers. A metal can only transfer by negative charges.
We call a semiconductor that uses mostly positive charge carriers (called holes) a p-type, and that which conducts electricity by negative
charge carriers (electrons) an n-type. These two ways to conduct allow engineers to create a solid material where charge can pass in only one
direction.
Engineers make a diode by placing p and n type semiconductors together. The n-type has a number of atoms that have an extra electron,
and these electrons are mobile. When the electrons move, they leave a fixed positive charge -- keep in mind, though, that the atoms themselves
aren’t mobile. The p-type semiconductor has acceptor atoms (fixed in the lattice) that contribute positive charge carriers called holes. When
this junction forms, positive holes are attracted to the negative electrons, which will flow across the pn junction. You would think these would
continue until the charge was even across both semiconductors, but the atoms fixed in place in the lattice prevent this. When a donor
contributes a mobile electron, it leaves a positively-charged atom fixed in place. This means that when the positive holes move into the n-type
semiconductor, they eventually meet a “wall” of fixed positive charges that repels them. Similarly, on the p-type side, each mobile hole leaves
behind a negatively-charged atom fixed forever in the lattice. This, of course, creates a wall of negative charges, which stops the mobile
electrons from moving further.
What Exactly Is a Hole?
We talk glibly of positive holes in a semiconductor, yet they exist only in the minds of electrical engineers. Holes are a fiction, albeit a
clever and useful one. It might seem like a hole is really a positron -- the anti-matter counterpart of an electron. It isn’t. Positrons exist only for
fractions of a second and are generated in huge particle accelerators; holes seemingly, in contrast, flourish in all semiconducting devices.
Instead, the notion of holes is a shorthand way to identify one of two different ways electrons can move in a semiconductor and thus conduct
charge. If a semiconductor conducts electricity mostly by negative charge carriers (an “electron” in electrical engineering parlance), it is called
n-type; if it seems to use positive carriers, nicknamed “holes,” it is called p-type.
A solid-state diode is made from a p-type and an n-type semiconductor. In the p-type the mobile carriers are thought of as a
positive charge called “holes.” In the n-type, the mobile charge carriers are negatively charged.
A diode is formed by the junction of p and n-type semiconductors. As shown in (a), we think of the p-type as having mobile
positive carriers and the n-type as having negative charge carriers. When the types are brought together (b), the negative and
positive charge carriers diffuse across the interface until an equilibrium is reached. If we hook up a battery (c) such that the negative
end is attached to the n-type conductor and the positive end to the p-type, the negative end of the battery provides a stream of
electrons to neutralize the positive charges fixed in the n-type semiconductor. This means that the positive holes can flow because the
wall of negative fixed charges has been neutralized. Once this has happened, the electrons can flow through the p-type. If we reverse
the battery (d), a stream of electrons flows toward the p-type end. Recall that there was a set of fixed negative charges left from the
mobile positive holes, which prevented electrons from moving through the p-type. The reversed battery, then, simply worsens the
situation, making that end even more negative and thus making it harder for any electrons to move. Thus current comes to a halt.
Engineers create n-type semiconductors by adding an atom like arsenic to the silicon lattice. An arsenic atom has five electrons,
rather than the four surrounding a silicon atom. This means that arsenic adds an extra electron that can become mobile to a
semiconductor. A p-type semiconductor is made by adding an atom like boron, which has three electrons. When added to silicon,
boron allows an electron to move from bond to bond. As shown in the next figure, this transfer of electrons can be thought of as a
positive “hole” carrying charge.
Engineers can control the type of conduction by replacing some of the silicon atoms with similar atoms. A silicon atom has four electrons
that it can share to form bonds. In a pure crystal of silicon, this allows it to bond with four other silicon atoms. If we replace some of these
silicon atoms with arsenic atoms -- called impurities, which is odd because we want them there! -- an extra electron appears. Arsenic has five
electrons to share. When we toss it into the silicon lattice, the arsenic shares four of those electrons with neighboring silicon atoms, forming
bonds, but the fifth electron is free to roam the crystal. Since the arsenic itself has a charge of +5, it balances that roaming negative charge.
Thus the crystal remains electrically neutral over all. At low temperatures, these electrons become bound to the arsenic atoms, but at room
temperature, thermal agitation shakes them off and lets them be free to roam. That extra negative charge conducts electricity throughout the
crystal. This would be called an n-type semiconductor -- the “n” stands, of course, for negative.
We could, instead, replace some silicon atoms in pure silicon with an element that has fewer than four electrons. Boron, for example, has
only three electrons to share. When substituted for a silicon atom, it cannot complete bonds to all four of the surrounding silicon atoms: It has
only three electrons and so forms three bonds. Electrons can jump from one of the neighboring, fully-bonded silicon atoms and completes this
bond, leaving of course an electron-deficient bond behind. Now, one could look at this type of motion as an electron moving, although it is
different than the electron movement in a p-type semiconductor. There an electron wanders from the crystal freely, but here an electron jumps
from bond to bond. Now one could look at this as a positive charge carrier that is jumping around the crystal: Every time an electron leaves a
bond and completes the fourth bond of the boron, it leaves a positive charge on the bond it left. Most importantly, though, these two
mechanisms of charge transfer move in opposite ways to an applied electrical field, just like we would expect positive and negative charge
carriers to move.
In a semiconductor, all charge is transferred by electrons, but it occurs in two different ways. Shown here is how an electron
jumps in a p-type semiconductor from bond to bond when an electric field is applied as indicated. Initially, (a) an impurity boron
atom has three complete bonds (two shared electrons) and a bond deficient by one electron. As shown in (b), this impure atom can
borrow an electron from a nearby bond. This borrowing can repeat itself across the crystal as shown in (c), (d), (e), and (f). Since the
electron-deficient bond has a positive charge, this appears to be a positive charge -- called a hole -- moving from left to right across
the crystal.